Background. Thyroid hormones are essential for brain development. We conducted a randomized, controlled trial with thyroxine (T4) supplementation in infants <30 weeks’ gestation and with the last neurodevelopmental follow-up moment at the age of 5.5 years. T4 supplementation was associated with improved outcome of infants <28 weeks’ gestation and worse outcome of infants of 29 weeks’ gestation. We studied gestational age–dependent effects of T4 supplementation at the mean age of 10.5 years in children participating in our randomized, controlled trial.
Methods. Questionnaires regarding school outcome, behavior, quality of life, motor problems, and parental stress were sent to the parents and children and their teachers at the same time point for all surviving children (9–12 years of age).
Results. Seventy-two percent of the families responded to our questionnaires. Nonrespondents had more sociodemographic risk factors and worse development until 5.5 years. At the mean age of 10.5 years, T4 supplementation was associated with better school outcome in those who were <27 weeks’ gestation and better motor outcome in those who were <28 weeks’ gestation, whereas the reverse was true for those who were born at 29 weeks’ gestation. No other gestational age–dependent outcomes were found.
Conclusions. Gestation-dependent effects of T4 supplementation remain stable over time. These effects do not prove beneficial effects of T4 in infants <28 weeks but should be the background for a new randomized, controlled trial with thyroid hormone in this age group.
- premature infants
- psychomotor development
- controlled trial
- school performance
- thyroid function
Thyroid hormones are essential for normal brain development.1 In preterm infants, plasma concentrations of the protein-bound thyroxine (T4) and triiodothyronine (T3), as well as the unbound free T4 (FT4) and T3, are lower in infants of lower gestational age.2–4 This is especially true in infants of <28 weeks’ gestation.5,6 In these infants, like in fetal life, inactivation of T4 is still a dominant metabolic pathway and results in high concentrations of sulfated thyroid hormones and reverse T3.5,6 Low concentrations of T4, T3, and FT4 in the neonatal period are associated with worse neurodevelopmental outcome up until the age of 9 years.7–9 Therefore, shortage of thyroid hormone could be one of the causes of impaired neurodevelopment that is often found in very preterm infants.10
To determine whether T4 supplementation can improve developmental outcome of infants, we conducted a randomized trial between 1990 and 1993. When the children were 2 and 5.5 years of age, we did not find any difference in outcome between the (complete) T4- and placebo-treated groups. However, at both ages T4 supplementation was associated with clinically and statistically improved outcomes in the subgroup of infants born at <28 weeks’ gestation, especially in those <27 weeks’ gestation.11,12 These infants have the lowest concentrations of (free) T4, accompanied by the above-mentioned immature metabolic pathways. On the other hand, T4 supplementation was associated with a worse outcome for infants born at 29 weeks’ gestation. This negative finding was not found to be related to high FT4 concentrations resulting from T4 supplementation.9 Also, all FT4 concentrations in the supplemented group remained within our in-house normal limits for FT4 in the neonatal period.
In the present study, we wanted to determine whether these gestation-dependent effects of T4 supplementation were still present when the children were 9 to 12 years old.
PATIENTS AND METHODS
All patients in this study participated in a randomized, controlled, double-blind trial of T4 supplementation.11 Two-hundred infants of 25 to 29 weeks’ gestation were included in this study before the 24th hour of life between January 1991 and July 1993. Details of the randomization process and study design are described elsewhere.11 T4 (or placebo) was administered during the first 6 weeks of life in a dose of 8 μg/kg birth weight per day. Thirty-five infants died in the neonatal period, 7 were withdrawn from the study, 1 child moved abroad, and the parents of 1 child refused to participate in additional studies when the child was 5 years old, resulting in a follow-up group of 156 children.
After informed consent was obtained for this part of the study, questionnaires were sent to the home addresses of the 156 children.
The child’s long-term outcome was assessed with questionnaires administered to parents and the children and their teachers. This was done at the same time for all children, allowing an age range of 9 to 12 years. Information on the child’s motor functioning was obtained from the parents, who used the Movement Assessment Battery for Children (ABC) Checklist.13 Behavioral-emotional functioning was assessed by using the Dutch version of the Child Behavior Checklist14 (CBCL), to be completed by the parents, and the Teacher Report Form, to be completed by the teacher. The child’s quality of life was assessed by using the Dutch TNO-AZL Children’s Quality of Life Questionnaire,15 consisting of a child and a parent version. Mothers completed the Dutch Nijmegen Parental Distress Index-Short Version.16 Additional information about the child’s school level, family demographics, and psychosocial care was obtained from the mother.
All statistical calculations were performed by using SPSS 10.1 (SPSS Inc, Chicago, IL). Univariate analyses were performed by using the Student’s t and χ2 tests for continuous and categorical data, respectively. For analyses of gestation-dependent effects of T4, multivariate analyses of variance and logistic-regression analyses were performed, using T4 × gestational age as an interaction term together with the following prerandomization covariates: gestational age, birth weight, birth weight class (appropriate for gestational age, small for gestational age, and very small for gestational age), maternal educational level, need for intubation at birth, and need for surfactant rescue therapy. When the interaction term was significantly associated with the outcome variable tested, treatment effect was studied separately in 4 gestational age groups: 25 + 26, 27, 28, and 29 weeks. Because of small numbers in these 4 groups, only univariate analyses were performed.
The parents and teachers of 113 (72%) of the 156 children who remained in the study responded positively and returned all or some of the questionnaires. The nonrespondents were more often born of nonwhite parents (with somewhat lower education level) and less often had a cerebral hemorrhage in the neonatal period. No other perinatal or sociodemographic differences were found between respondents and nonrespondents. There were also differences regarding developmental outcomes at 2 and 5.5 years of age. Nonrespondents had significantly lower developmental outcome at 2 years, lower IQ scores at 5.5 years, and significantly higher CBCL behavioral problem scores at 5.5 years (Table 1).
Table 2 shows the perinatal, sociodemographic, and developmental characteristics of the 58 T4-treated and 55 placebo-treated children of the respondent group. There were no significant differences on each of these domains.
Similar to our studies at 6, 12, and 24 months and 5.5 years, at the mean age of 10.5 years (range: 9–12 years) there were no differences in outcomes regarding school outcome, behavior, quality of life, motor functioning, and parental stress between the T4- and placebo-treated groups (Tables 3 and 4). Special schooling was needed in 19% of T4-treated children and 22% of placebo-treated children, whereas such schooling is needed for only 6% of the general Dutch population in this age group. In a nationwide study of school outcomes of 9-year-old children born in 1983 at <32 weeks’ gestational age and/or with a birth weight of <1500 g, 27% needed special schooling.17
To establish whether there were gestation-dependent effects of T4, comparable to our results at the ages of 2 and 5.5 years, we first performed multivariate analysis with T4 × gestational age as an interaction term. There was a significant interaction effect regarding the need for special education (P = .006) and motor function (P = .023), whereas no such interaction effect was found for behavioral, parental stress, and quality-of-life outcomes.
Figure 1A shows the percentage of children in special schools for T4- and placebo-treated children separately per gestational week. In the group of 25 and 26 weeks’ gestation, 5 times as many placebo-treated than T4-treated children needed special education; also, in the 27 and 28 weeks’ gestation groups, T4-treated children performed somewhat better than placebo-treated children, whereas in the group of children born at 29 weeks’ gestation, the reverse is true, with more T4-treated than placebo-treated children in special education. Reasons for special education included motor handicaps and learning and behavioral problems; 1 child visited a school for children with hearing problems. There were no children with severe visual deficits. Groups were too small to identify possible differences in reasons for special education among the various groups.
In Fig. 1B a comparable pattern is seen with respect to motor outcome. Motor scores were almost 3 times higher (worse) in placebo-treated children who were born at <28 weeks’ gestation compared with T4-treated children of the same gestational age groups. Again, in children born at 29 weeks’ gestation, T4-treated children performed worse than placebo-treated children, although in this gestational age group both study groups had better scores than were found in the groups of lower gestational ages.
Although several studies have demonstrated correlations between low thyroid-hormone concentrations in the neonatal period and neurodevelopmental outcome of (very) preterm infants, only a few randomized trials with thyroid hormone have been undertaken in the surfactant era.11,18–20 Moreover, most of these trials have focused on clinical, instead of neurodevelopmental, outcomes.18–20 Our trial was aimed primarily at improvement of neurodevelopmental outcome at 2 years of age. Based on the available literature on hypothyroxinemia published before 1989, we chose to include infants of <30 weeks’ gestational age. At present, there is more knowledge on FT4 concentrations, which are more important for the biological effect than bound-T4 concentrations.21 FT4 concentrations are decreased to a much lesser extent than bound-T4 concentrations,3–5 and thyroid-hormone deficiency in preterm infants seems to be a subtle problem that increases in severity in more immature infants.5,22
In view of the results discussed above, our present results can be of importance for additional studies.
Despite the limited follow-up number of 72%, our results are stable in time, showing also at the age of 9 to 12 years possible beneficial effects of T4 supplementation in infants of <28 weeks’ gestation and reversal of this effect in infants of 29 weeks’ gestation.
The need for special education seemed to be the strongest outcome parameter in this phase of the study. It is determined by the sum of all developmental problems that preterm infants may show and includes neuromotor, behavioral, cognitive, and sensory handicaps. Within the current material no additional specification of effect can be made toward a certain area in the brain; however, comparable with our results at 5.5. years, neuromotor problems seem to be occurring less after T4 supplementation.
The gestation-dependent positive effects of T4 in infants of <28 weeks’ gestation could have several explanations. FT4 concentrations in the more immature infants might indeed be too low3–5 to ensure normal brain development. It could also be caused by thyroid-hormone–dependent maturational processes in the brain, which may especially take place in the window of time before 29 weeks. Also, in the timing of restoring maternofetal iodine deficiency, there is a similar window of time in the second trimester, during which iodine supplementation can improve head circumference and developmental outcome but after which iodine supplementation is unsuccessful.23 Alternatively, thyroid hormones may be especially effective in preventing or repairing ischemic brain damage, which occurs predominantly in the most immature infants.24
The negative findings in the group of children born at 29 weeks’ gestation are more difficult to explain. In an earlier analysis,9 we did not find a relation with higher FT4 concentrations and worse outcome. In our trial, however, we administered T4 as a bolus, and it might be more physiologic to supplement in a continuous way, as was done by Biswas et al.20 We started our supplementation 24 hours after birth, during the postnatal T4 surge that occurs normally after birth; therefore, we do not have reliable information on the degree of hypothyroxinemia in the T4-supplemented group and might have included infants who did not have hypothyroxinemia at all, especially in the group of infants born at 29 weeks’ gestation.5
Of course, it is also conceivable that our findings are the result of chance. Indeed, they were found by posthoc analysis, and group differences, which remained stable in time, could also be the result of other biological, environmental, or genetic risk factors.
Only a new trial with T4 supplementation in infants of <28 weeks’ gestation can tell whether our findings are based on a positive biological effect of T4. In light of the often-impaired neurodevelopmental outcome of NICU graduates, it would seem to be a useful trial to undertake.
- Accepted June 13, 2005.
- Address correspondence to Aleid G. van Wassenaer, MD, PhD, Department of Neonatology, Emma Children’s Hospital Academic Medical Center, H3N, Meibergdreef 9, 1105 AZ Amsterdam, Netherlands. E-mail;
No conflict of interest declared.
- ↵Bernal J, Nunez J. Thyroid hormones and brain development. Eur J Endocrinol.1995;133 :390– 398
- ↵Biswas S, Buffery J, Enoch H, Bland JM, Walters D, Markiewicz M. A longitudinal assessment of thyroid hormone concentrations in preterm infants younger than 30 weeks’ gestation during the first 2 weeks of life and their relationship to outcome. Pediatrics.2002;109 :222– 227
- Lucas A, Morley R, Fewtrell MS. Low triiodothyronine concentration in preterm infants and subsequent intelligence quotient (IQ) at 8 year follow up. BMJ.1996;312 :1132– 1133; discussion 1133–1134
- ↵Van Wassenaer AG, Briët JM, Van Baar AL, et al. Free thyroxine levels during the first weeks of life and neurodevelopmental outcome until the age of 5 years in very preterm infants. Pediatrics.2002;109 :534– 539
- ↵Perlman JM. Neurobehavioral deficits in premature graduates of intensive care—potential medical and neonatal environmental risk factors. Pediatrics.2001;108 :1339– 1348
- ↵Briet JM, Van Wassenaer AG, Dekker FW, De Vijlder JJM, Van Baar AL, Kok JH. Neonatal thyroxine supplementation in very preterm children: developmental outcome evaluated at early school age. Pediatrics.2001;107 :712– 718
- ↵Hendersen SE, Sugden DA. Movement Assessment Battery for Children-Checklist. Lisse, Netherlands: Swets and Zeitlinger; 1998
- ↵Verhulst FC, Van den Ende J, Koot HM. Manual for the CBCL 4-18 [in Dutch]. Rotterdam, Netherlands: Department of Child en Adolescent Psychiatry, Sophia Children’s Hospital/Academic Hospital Rotterdam/Erasmus University; 1997
- ↵Verrips EGH, Vogels TGC, Koopman HM, et al. Measuring health-related quality of life in a child population. Eur J Public Health.1999;9 :188– 193
- ↵Abidin RR, De Brock AJLL, Gerrits JRM, Vermulst AA. NOSI: Nijmeegse Ouderlijke Stress Index. Lisse, Netherlands: Swets and Zeitlinger; 1992
- ↵Hille ET, den Ouden AL, Bauer L, van den Oudenrijn C, Brand R, Verloove-Vanhorick SP. School performance at nine years of age in very premature and very low birth weight infants: perinatal risk factors and predictors at five years of age. Collaborative Project on Preterm and Small for Gestational Age (POPS) Infants in the Netherlands. J Pediatr.1994;125 :426– 434
- ↵Biswas S, Buffery J, Enoch H, Bland M, Markiewicz M, Walters D. Pulmonary effects of triiodothyronine and hydrocortisone supplementation in preterm infants less than 30 weeks gestation: results of the THORN Trial—Thyroid Hormone Replacement in Neonates. Pediatr Res.2003;53 :48– 56
- ↵Van der Meulen BF, Smrkovsky M. De Bayley Ontwikkelingsschalen (BOS 2-30). Lisse, Netherlands: Swets and Zeitlinger; 1983
- Copyright © 2005 by the American Academy of Pediatrics