Unidad de Endocrinología Molecular Instituto de Investigaciones Biomédicas 28029-Madrid, Spain Susana Ares Servicio de Neonatología Hospital Infantil La Paz 28046-Madrid, Spain
Dr. Delbert A. Fisher introduces his careful editorial on this subject, in the June issue of JCEM (1), with a summary of known alterations of thyroid function in premature newborns that supports the idea that "for premature infants the stresses of extrauterine transition are superimposed on an immature thyroid axis ... Thus the premature infant in the extrauterine environment, like the intrauterine fetus of comparable gestational age (GA), is relatively hypothyroxinemic with normal-to-low levels of serum TSH and T3 concentrations ... The prevailing view has been that the hypothyroxinemia of prematurity is physiologic, not requiring hormone supplementation."
After reviewing recent data of several lines of evidence suggesting that hypothyroxinemia during this premature period of development has a negative impact on later development, especially of the brain, the author concludes that the prevailing view quoted above is still valid.
We wish to question the statement that findings in the preterm neonates
in the extra uterine environment can be extrapolated to the
intrauterine fetus of comparable gestational age (GA). In the study by
Ares et al. (2) appearing in the same issue, the free T4
(FT4) concentrations found in the premature infants were much lower
than those reported for the intrauterine fetus at a comparable
developmental stage by Thorpe-Beeston et al. (3), as shown
in the upper panel of Fig. 1.Although Ares et al. (2) also indicated that the circulating
levels of TSH of both preterm and term neonates were below the
confidence intervals for the fetal TSH levels reported by
Thorpe-Beeston et al. (3), the actual data were not shown.
They now appear in the bottom panel of Fig. 1
. As may be seen, except
for the values at the earliest postmenstrual age, which were obtained
within the first week after birth, and therefore likely to include the
postnatal TSH surge, all other TSH values were lower than reported
in utero. As already stressed by Thorpe-Beeston et al. (3),
TSH is positively correlated with FT4 in the fetus, independently of
GA, and continues to increase until term, when adult FT4 levels have
been reached. These intrauterine TSH levels are much higher than those
of young term infants after the postnatal TSH surge is over, or those
of euthyroid adults.
|
In the editorial (1), the very low FT4 levels found in premature
neonates of the Madrid study (2) are questioned because they were not
obtained by an equilibrium dialysis method. Neither were those reported
by Thorpe-Beeston et al. (3), and we believe the comparison
appearing in Fig. 1 is valid, as the values obtained in the older
fetuses (3) and in our term neonates (2) were comparable. Although the
reference values recently reported in premature infants from California
using a direct dialysis method are higher (4) than those of the Madrid
study or those reported by Thorpe-Beeston for older fetuses (3), it
should be pointed out that the reference samples submitted to the
dialysis method were obtained quite soon after birth (after
approximately 3.5 ± 2.5 days), when the surge in FT4 following
parturition might still affect the findings, a possibility strongly
supported by the very high TSH found in the same samples. The low FT4
levels from the Madrid study are influenced both by postmenstrual age
and iodine intake, independently. By multiple regression analysis of
the standardized variables (z-scores) it was found that two
thirds of FT4 value was related to age and one third to the iodine
intake. Even if the lowering effect of the iodine deficiency were taken
into account, circulating FT4 of the premature infants remained lower
than that of fetuses of comparable age. The marked differences in the
circulating TSH of fetuses and preterm infants of comparable GA are not
easily accounted for on the basis of methodological differences.
The other important point reviewed in Dr. Fishers editorial is whether or not the hypothyroxinemic premature infants would benefit from hormone replacement therapy. Evidence reviewed by Fisher (1) and Ares et al. (2) suggests that the critical period of central nervous system dependency on thyroid hormones in humans extends to intrauterine fetal life and to the extra uterine premature infant. Two double-blind studies have addressed this question, namely, those reported by Chowdry et al. (5) and van Wassenaer et al. (6). The first one was limited to a small number of infants and did not detect any beneficial effect of treatment with thyroxine (1015 µg/kg per day) on neurodevelopment, as measured at 1 yr of age in 5 T4-treated as compared with 3 placebo-treated infants. The second study is based on 100 premature babies of less than 30 GA, who were treated daily with 8 µg/kg of birth weight for 6 weeks, and 100 placebo-treated controls. The latter study showed several important points, the first one being that treatment was not harmful. Although neurodevelopment, as measured at 2 yr of age, was not improved for the cohort of treated infants taken as a whole, it had a clear beneficial effect for those born at less than 27 weeks GA, whose IQ scored 18 points higher than that of placebo-treated infants (93 ± 26, compared with 75 ± 23, P < 0.01). These subgroups consisted of 13 T4-treated versus 18 placebo-treated infants. Van Wassenaer et al. (7) have also recently reported that mortality and morbidity of the preterm neonates born from mothers who had not received antenatal steroid treatment (32 T4-treated vs. 35 placebo-treated neonates) were significantly improved: 14 (40%) died in the placebo-treated group, compared with only 5 (16%) in the T4-treated group. Decreased mortality of early preterm neonates had also been reported by Schönberger et al. (8), who treated 45 infants daily with 25 µg + 5 µg T3 given orally: 7 (3%) died, as compared with 16 out of 55 (29%) untreated infants. They later reported a 9.5% mortality rate for the successive 126 prematures, born weighing less than 2200 g or before 37 weeks gestation, who received the treatment (10). Considering the publication dates, it is likely that the premature infants from these studies (9, 10) were born to mothers who had not received antenatal steroid treatment. Although the numbers are small, results appear to be of sufficient clinical relevance to encourage more attempts at treating the neonatal hypothyroxinemia of prematurity.
We again wish to point out that a low FT4 level is not the only main difference between the thyroid hormone status of the preterm infant and a fetus in utero at a comparable developmental age. None of the treatment schedules used so far have taken into consideration that TSH is higher in the fetus. The doses of T4 that have been used actually depressed circulating TSH even below the levels found in nontreated preterm infants (6). Possibly as a result of this, the thyroidal secretion of T3 would decrease. This would occur during a developmental period, when type I 5' iodothyronine deiodinases (5'D-I) are still low. Treatment with T4 also resulted in low circulating T3, and this would further compromise the T3-dependent expression of 5'D-I. Treatment schedules that would supply the thyroid hormone requirements of the premature infant without delaying thyroidal and 5'D-I maturation might improve the encouraging developmental results obtained so far.
In summary, we do not question that the hypothyroxinemia of the premature infant reflects an inherent hypothalamic-pituitary-thyroid immaturity. We do disagree, however, with the idea that this is a physiological condition for the premature neonate, because the intrauterine environment to which it would be exposed if still in utero would avoid the hypothyroxinemic state and, as a result of the high circulating TSH levels, would promote thyroidal maturation. The consequences of these contrasting points of view are quite different, with the first discouraging, and the latter encouraging therapeutic measures to avoid this period of hypothyroxinemia, which has been shown to be causally related to poor neurodevelopment (10, 11).
Footnotes
Received October 16, 1997. Address correspondence to: Gabriella Morreale de Escobar, Instituto de Investigaciones Biomédicas, Facultad de Medicina, UAM, Arzobispo Marcillo 4, 28029 Madrid, Spain.
References