Thyroid hormone is known to regulate neurodevelopment, probably from early fetal life onwards (1). Thyroid hormone deficiency can cause long term morbidity in terms of behaviour, locomotor ability, cognition and hearing ability, if the onset is early in development (2).
Since the introduction of neonatal screening for congenital hypothyroidism in the 1970's it became clear that preterm infants have lower plasma concentrations of (free)T4 and (free)T3 than full term infants of the same postnatal age and this has raised an ongoing discussion on the need for thyroid hormone supplementation in preterm infants in order to improve clinical and neurodevelopmental outcome.
The thyroid state of the preterm infant can be viewed as a subtle form of hypothyroidism or hypothyroxinemia. As in maternal hypothyroxinemia during the first half of pregnancy (3), a low supply of thyroid hormone to the fetus could be harmful for brain maturation during its early development (4), and this phenomenon must be studied carefully not only in a descriptive manner, but also in the light of a possible need for thyroid hormone supplementation.
 
Fetal thyroid state
 
By 7 weeks of gestation, before the onset of fetal thyroid functioning, thyroid hormone has been demonstrated in fetal fluids, including serum, FT4 concentrations being at least one third of the mother's FT4 concentration (5). Also thyroid hormone receptors and the deiodinating enzymes are present in human fetal cerebral cortex by that time (1). Nuclear thyroid hormone receptors occupied by bioactive T3 have been found in human brain and lung tissue in the 9th week of fetal life. Although, the hypothalamus and pituitary start to synthesize hormones by 12 weeks, significant thyroid hormone production does not occur before the 20th week of gestation (6). Before mid-gestation, materno fetal transfer of T4 is therefore pivotal for the fetal thyroid hormone status.
Fetal thyroid function and the hypothalamic-pituitary-thyroid axis continue to mature throughout pregnancy; serum levels of TT4, FT4, thyroglobulin, and TSH increase until the end of pregnancy (7). Serum levels of TT4 increase from about 5 nmol/l at 12 weeks gestation to about 120 nmol/l at term, while the increase of serum TT3 is much less: from 0.5 nmol/l at 12 weeks to about 1.5 nmol/l at term (5,7). Serum FT4 in cord blood seems to increase from about 5 pmol/ at 12 weeks gestation to about 20 pmol/l at term (5,7).
During fetal life, the concentration of TT3 is tightly controlled in the tissues. The already abundantly present T4 is preferably converted by type III deiodinase to rT3, which is present in high concentrations during fetal life and only decreases in the last weeks, while T3 is readily converted to diiodothyronine. Also, sulfatation by hepatic sulfotransferase enzymes to the inactive sulfated metabolites T4 sulfate, T3 sulfate, and rT3 sulfate is an inactivating metabolic pathway in fetal life (8).
Presumably, when increases in local intracellular T3 concentrations are needed for thyroid hormone-dependent maturational processes, local type II deiodinase increases, converting T4 to T3, whereas type III deiodinase activity decreases, favoring intracellular T3 accumulation. In this respect, the plasma T4 concentration is far more important than the plasma T3 concentration. The local concentration of thyroid hormone receptors and possibly mechanisms regulating T4 uptake in the cells also play a role in this ontogenetically programmed production and action of T3 (1).
The described regulatory mechanisms are also important in protection against thyroid dysfunction. Thus, also in human fetal brain, like in the rat brain, type II deiodinase, was found to increase in response to plasma T4 decrease (9), but the onset of this regulatory mechanism was only found at mid-gestation.
  
Postnatal function in the preterm infant
 
After preterm birth, TT4 and TT3 levels remain lower than in term born infants during the first weeks (10). There is an obtunded TSH peak immediately after birth, while it remains below 20 mU/l, being the cut-off point for congenital hypothyroidism, in the period of low TT4. This period during which total and free T4 (and T3) levels are low is generally referred to as transient hypothyroxinemia of the preterm infant.
 

Table 1. Factors that influence (very) preterm thyroid function     Immaturity of the hypothalamic-pituitary-thyroid axis Immature thyroidal capacity to concentrate iodine and synthesize and iodinate thyroglobulin Increase of thyroid hormone needs for example for thermogenesis, heart function, skeletal muscles etc Sudden interruption of materno-fetal transfer of T4 Immaturity of thyroid hormone metabolism, causing low T3 and high rT3 and sulfated iodothyronines Effects of neonatal disease (non-thyroidal illness) Insufficient iodine supply Iodine excess (iodine containing antiseptics and radio opaque agents)

Table1
 

In infants of less than 30 weeks gestational age, TT4 concentrations are about 60 nmol/l in the first week of life (11), while in term infants TT4 concentrations are generally 4 times higher. Postnatal free thyroid hormone concentrations are also lower the earlier in gestation the infant is born. FT4 concentrations are about 2-fold lower in very preterm infants as compared with term infants of 1 week of age. (12, see also the figure).
 

Figure:
Mean Free T4 concentrations in the first 5 weeks after birth, as measured by a two-step RIA in infants of 25-28 weeks gestation (…..), infants of 28-30 weeks gestation (-----) and of term infants as estimated from literature (->), (ref. 12,14,25)


The average of five FT4 measurements drawn between day 3 and 28 in infants <30 weeks gestation, by a two-step RIA, was found to be between 10.1 and 21.1 pmol/l (13). TSH has a variable time course, but comes down to about 2-4 mU/l by 4 weeks after birth (14). The postnatal time course of T3 in preterm infants misses the sharp peak after birth and only slowly rises to term values in the course of 6-8 weeks (10,14).

Preterm infants are at risk for neurodevelopmental impairments. The more preterm the infant is born, the higher the risk of neurological impairments. These concern speech, language, behaviour and learning, and in serious cases overt mental retardation may occur. The neuromotor deficiencies vary from clumsiness to disabling cerebral palsy. Visual and hearing impairments are also frequent.
Four studies (Table 2 ) show an association between low thyroid hormone levels in the first weeks of life and worse developmental outcome.

Table 2. Retrospective cohort studies on the association between neonatal plasma (F)T4 and T3 levels in preterm infants and later development.             Source   Findings Lucas A, et al. 1988, 1996 (15,16) N=236   Low neonatal TT3 is associated with lower IQ both at 18 months and 8 years of age       Meijer WJ et al 1992; den Ouden AL et al 1996 (4,17) n=563   TT4 at day 7 of life is associated with developmental delay at 2 years of age and school problems and minor neurological dysfunction at 9 years of age       Reuss et al 1996   A TT4 at <day 7 of life of >2.6 SD below the test mean is associated with an increased risk of disabling cerebral(18) N=463 palsy and a 7 points IQ reduction       Van Wassenaer et al 2002 (13) N=75   Low FT4 during first 4 weeks of life is associated with worse neurodevelopmental outcome at 2 and 5 years

Table2


Of course they do not provide evidence that preterms infants should be treated with thyroid hormone. Low thyroid hormone levels are also associated with higher mortality and more respiratory disease (13,14,19), a higher incidence of cerebral hemorrhage (19) and ischaemic lesions (20).
Only randomized clinical trials, testing the effect of thyroid hormone treatment in preterm infants, can untangle the complicated relationships between thyroid hormone levels, gestational age, morbidity and neurodevelopment.

Studies with thyroid hormone administration in preterm infants

Between 1997 and 2003 four randomized, double-blind trials were published (21,22, 24,25), see Table 3, with different treatment protocols and endpoints.

Table 3. Summary of four randomized double-blind T4 or T3 treatment studies in infants <32 weeks gestational age.
study	Intervention	Endpoint	no. (T vs co) at assessment of endpoint gestational age (Ga)	main results
Vanhole, et al 1997(21)	T4, daily i.v. bolus, 20 mg/kg, d1-14	Endocrine and Clinical (also neuro-development at 7 mo	17 versus 17 Ga <31 wk	No difference in clinical outcome and development
Van Wassenaer et al, 1997 Briet et al, 2001(22,23)	T4, daily bolus, first 2-3 wks iv, later orally; 8 mg/kg, d 1-42	Neuro-development at 24 mo; and outcome at 5.7 yr	82 versus 75 Ga <30 wk	No difference in total groups. Subgroup analyses: at 2 and 5yrs better outcome with T4, if Ga <27-29 wks
Smith et al, 2000(24)	T4, bolus, start iv: 10 mg/kg; then orally: 20 mg/kg, d 2-21	Chronic lung disease Need for supplemental oxygen at day 28	29 versus 18 Ga <32 wk	No effect
Biswas, et al 2003(25)	T3, continuous iv, 6 mg/kg/d plus hydrocortisone 1 mg/kg/d; d 1-7	Death or ventilator dependence at d7	125 versus 128 Ga <30 wk	No difference in adverse outcome at d7

Three of the studies (21,24,25) chose short term clinical outcome as primary end point, in one of them T3 was administered instead of T4 (25). In none of the studies mortality or morbidity was significantly influenced by thyroid hormone treatment.
In only one study neurodevelopmental outcome was chosen as primary end point (22,23); children were assessed five times between corrected term age and 5.7 years. Neurodevelopmental outcome was similar in both groups at all time points. However, both at two years and at 5.7 years, post-hoc subgroup analyses revealed a gestational age-dependent effect of T4. T4-treated infants of <29 weeks gestation had a better neurodevelopmental outcome, but for T4-treated infants of 29-30 weeks the reverse was true (23). The latter possible harmful effect of T4 was not related to high FT4 concentrations (13).
Taken together, none of these studies provides evidence for the need of thyroid hormone supplementation in very preterm infants and therefore the current advice is to not supplement low thyroid hormone concentrations in these infants, unless accompanied by elevated TSH (26,27) .

Conclusions and Recommendations for further studies

Until now, a low T4 with TSH of less than 20 mU/L has been used in the definition of transient hypothyroxinemia, with cut-off values for TT4 that vary between the different authors from 40 to 100 nmol/l. TBG concentrations are also low, however, and therefore FT4 may even be high, when TT4 is in the low range. In our material, we found that TT4 concentrations of <60 nmol/l are accompanied by FT4 concentrations between 5.2 and 16.6 pmol/l with 60% of FT4 values below 10 pmol/l (unpublished observation). It is therefore necessary to include the FT4 concentration in the definition of transient hypothyroxinemia. The normal range for FT4 concentrations in preterm infants should be established, but this can only be done if developmental outcome of these infants is part of these studies. Normal ranges of FT4 should be known for any specific type of assay, as FT4 is reported to be higher by dialysis method than by other methods (28). Because of the strong association with gestational age and birthweight, normal values should be established per gestational age or birthweight group .

Using different treatment protocols in each of the randomized controled trials, researchers have not been able to demonstrate a positive effect of T4 and/or T3 treatment on clinical outcome. Presumably, improvement of clinical outcome should not be the aim of studies but rather of neuro-developmental outcome when thyroid hormone supplementation studies are designed. Our own study is the only study of this type (22). The results of the post-hoc subgroup analyses of our study seem to show that T4 supplementation may be beneficial in infants of less than 28-29 weeks of gestation (22,23)
Therefore, a new randomised controlled trial with T4 in a selected patient group of infants of <29 weeks gestation, who also have a FT4 measurement in the low range during the first 3 days of life, appears to be a logic next step. Whether the thyroid status of the mother contributes to that of her child(ren) is a question that also has not been answered so far .

 

REFERENCES