Home > Health & Research > Health Education Campaigns & Programs > Cochrane Neonatal Review > Thyroid hormones for preventing neurodevelopmental impairment in preterm infants

Thyroid hormones for preventing neurodevelopmental impairment in preterm infants

Skip sharing on social media links
Share this:

Authors

Osborn DA

Background - Methods - Results - Characteristics of Included Studies - References - Data Tables & Graphs


Dates

Date edited: 23/08/2001
Date of last substantive update: 12/06/2001
Date of last minor update: / /
Date next stage expected / /
Protocol first published: Issue 2, 1998
Review first published: Issue 2, 1999

Contact reviewer

Dr David A Osborn

Staff Specialist
Neonatal Medicine
Royal Prince Alfred Hospital
Missenden Rd
Camperdown
NSW AUSTRALIA
2050
Telephone 1: 61 2 95158760
Telephone 2: 61 2 95156111
Facsimile: 61 2 95504375

E-mail: davido@peri.rpa.cs.nsw.gov.au

Secondary address:
37 Station St
Pymble
NSW AUSTRALIA
2073
Telephone: 61 2 91441050

Contribution of reviewers

Intramural sources of support

  • None noted.

Extramural sources of support

  • None noted.

What's new

This review updates an existing review: "Osborn DA. Thyroid hormone for preventing neurodevelopmental impairment in preterm infants (Cochrane Review). In: The Cochrane Library, Issue 2, 2001. Oxford: Update Software. (A substantive amendment to this systematic review was last made on 11 February 1999). The review incorporates an additional study of thyroxine in preterm infants (Smith 2000) and the 5 year follow up from a previously included study (van Wassenaer 1997). These new data do not change the conclusion that there is no evidence from randomized trials to date that unselected preterm infants benefit from treatment with thyroid hormones.

Dates

Date review re-formatted: 15/09/1999
Date new studies sought but none found: / /
Date new studies found but not yet included/excluded: / /
Date new studies found and included/excluded: / /
Date reviewers' conclusions section amended: / /
Date comment/criticism added: / /
Date response to comment/criticisms added: / /

Synopsis

No evidence from trials that thyroid hormone therapy is effective in preventing problems such as respiratory distress syndrome in preterm babies.

Thyroid hormones are needed for the normal growth and maturity of the central nervous system, as well as the heart and lungs. Children born without sufficient thyroid hormones can develop serious mental retardation. It is believed that low levels of thyroid hormones in the first few weeks of life (transient hypothyroxinemia) in preterm babies born before 34 weeks may cause this abnormal development. The review of trials found no evidence that using thyroid hormones in preterm babies is effective in reducing the risk of problems caused by insufficient thyroid hormones.

Abstract

Background

Observational studies have shown an association between transiently low thyroid hormone levels in preterm infants in the first weeks of life (transient hypothyroxinemia) and abnormal neurodevelopmental outcome. Thyroid hormone therapy might prevent this morbidity.

Objectives

To assess whether thyroid hormone therapy in preterm infants without congenital hypothyroidism results in clinically important changes in neonatal and long term outcomes in terms of benefits and harms.

Search strategy

The standard search strategy of the Neonatal Review Group was used. This included searches of the Oxford Database of Perinatal Trials, Cochrane Controlled Trials Register, MEDLINE, previous reviews including cross references, abstracts, conferences, symposia proceedings, expert informants and journal handsearching in the English language.

Selection criteria

All trials using random or quasi-random patient allocation, in which thyroid hormone therapy (either treatment or prophylaxis) was compared to control in premature infants.

Data collection & analysis

Primary clinical outcomes included measures of neurodevelopmental outcome and mortality. Assessment of trial quality, data extraction and synthesis of data, using relative risk (RR) and weighted mean difference (WMD), were performed using standard methods of the Cochrane Collaboration and its Neonatal Review Group.

Main results

Nine studies were identified that compared thyroid hormone treatment to control. Four randomized (Chowdhry 1984, van Wassenaer 1997; Vanhole 1997; Smith 2000) and one quasi-randomized study (Amato 1989) met inclusion criteria. All studies enrolled preterm infants < 32 weeks gestation, but used different timing, dose and duration of treatment with thyroid hormones. Four studies used thyroxine, whereas Amato 1989 used triiodothyronine. Only two studies with neurodevelopmental follow-up were of good methodology (van Wassenaer 1997, Vanhole 1997). All studies were of small size with the largest, van Wassenaer 1997, enrolling 200 infants.

Meta-analysis of five studies found no significant difference in mortality to discharge (typical RR 0.70, 95% CI 0.42, 1.17) in infants who received thyroid hormone treatment compared to controls. Meta-analysis of two studies (van Wassenaer 1997; Vanhole 1997) found no significant difference in the Bayley MDI or PDI performed at 7-12 months. van Wassenaer 1997 found no significant differences in the Bayley MDI and PDI at 24 months, incidence of cerebral palsy (RR 0.72, 95% CI 0.28, 1.84), death and cerebral palsy (RR 0.70, 95% CI 0.43, 1.14) or the RAKIT IQ score (WMD -2.10, 95% CI -7.91, 3.71) at 5.7 years of age. Fraction of inspired oxygen was lower in infants receiving triiodothyronine in one small quasi-randomized study, but not in infants receiving thyroxine in a randomized study. Duration of mechanical ventilation and incidence of chronic lung disease were not reduced in infants receiving early thyroid hormone therapy.

Reviewers' conclusions

This review does not support the use of thyroid hormones in preterm infants to reduce neonatal mortality, improve neurodevelopmental outcome or to reduce the severity of respiratory distress syndrome. An analyses of data from one study (van Wassenaer 1997) which showed benefits in infants 24-25 weeks gestation was not prespecified and should be treated with caution.

The small number of infants included in trials incorporated in this review limits the power of the meta-analysis to detect clinically important differences in neonatal outcomes.

Future trials are warranted and should be of sufficient size to detect clinically important differences in neurodevelopmental outcomes. They should consider enrolling those infants most likely to benefit from thyroid hormone treatment such as infants born at less than 27 weeks gestation.

[top]

Background

In preterm infants, values for serum thyroxine and free thyroxine in the first days of life vary directly with gestation (Rooman 1996, Reuss 1997). However, unlike term infants the concentrations of thyroxine and free thyroxine decrease to reach a nadir between day 10 and 14 of life that is more severe at lower gestations and birth weights (Frank 1996, Rooman 1996). Reuss 1997 found the incidence of infants with severely depressed thyroxine values (below 4 micrograms/dL) ranged from 40% at 23 weeks gestation to 10.2% at 28 weeks gestation. Furthermore, the levels of thyroxine and free thyroxine found in premature infants are lower than those seen in the normal fetus at similar gestational ages (Ballabio 1989, Thorpe-Beetson 1991, Radunovic 1991).

This period of low thyroid hormone levels in infants born prematurely has been termed transient hypothyroxinemia of prematurity. Suggestions as to the cause of these low thyroid hormone values include an immaturity of the hypothalamic-pituitary-thyroid axis in the extremely premature infant or a form of "nonthyroidal illness" (the "sick euthyroid syndrome") reflecting the infant's response to severe illness (monograph review by Paneth 1998).

Transient hypothyroidism (low thyroxine, high thyrotropin concentration) also occurs in up to 5% of neonates admitted to neonatal intensive care (Rooman 1996), 0.4% of infants with birth weights < 1500g (Frank 1996) and 1.8% of infants born < 29 weeks gestation (Reuss 1997). There is evidence to suggest that this may be due to exposure to iodine-containing antiseptics used in neonatal care (Linder 1997).

Thyroid hormones are necessary for the normal growth and maturation of the central nervous system (monograph reviews by Porterfield 1993 and Bernal 1995). Congenital hypothyroidism is strongly associated with abnormal neurodevelopment. Neurological cretinism in its severest form is characterised by profound mental retardation, deafmutism, spastic diplegia and squint (Porterfield 1993). Even children who receive early thyroid hormone replacement therapy for congenital hypothyroidism have motor and cognitive deficits that persist to late childhood (Kooistra 1994).

Infants born prematurely also have a high incidence of motor and cognitive deficits that are worse at lower gestations (Lorenz 1998). The question is whether the transient decrease in serum concentration of free thyroxine that occurs in some preterm infants during the first weeks of life is harmful.

Three cohort studies (Lucas 1988 and 1996, Miejer 1992 and den Ouden 1996, and Reuss 1996) have documented an association between low thyroid hormone levels (triiodothyronine or thyroxine) in the first weeks of life and abnormal neurodevelopmental outcome. All three cohorts documented a measure of abnormal mental development in children who had low neonatal thyroid hormone levels. One study (Reuss 1996) found a 4.4 fold increase in risk of disabling cerebral palsy at 2 years. Den Ouden 1996 in the same cohort found children who had low neonatal thyroxine levels to have an increased risk of school failure at 9 years. The associations in the cohorts persisted despite correction for potential confounders including gestation, measures of fetal growth (either birth weight or presence of growth restriction) and, in some studies, many factors relating to severity of illness in preterm infants and independent risk factors for abnormal neurodevelopmental outcome.

Whether transient hypothyroxinemia of prematurity is a causative factor for abnormal neurodevelopment is uncertain. Transient hypothyroxinemia of prematurity is strongly correlated with low gestation and is more frequent in infants with respiratory distress syndrome and in ventilated infants (Franklin 1986). It may be that transient hypothyroxinemia of prematurity is secondary to low gestation and/or illness severity in preterm infants. It is possible that other factors which are responsible for the abnormal neurodevelopmental outcome in preterm infants have not been taken into account adequately in the cohort studies.

Thyroid hormones have also been demonstrated to have a synergistic effect in stimulating lung maturation and surfactant production in animal models (Liggins 1988). Although antenatal TRH administered to mothers to stimulate fetal thyroid hormone production has been shown to have limited effect (Cochrane review: Crowther 1998), it has been suggested that thyroid hormones administered shortly after birth may also reduce the severity of respiratory distress syndrome (Amato 1988). No evidence of an effect of thyroid hormones administered post-natally on the severity of respiratory distress syndrome in an animal model could be found in the literature.

This review examines the evidence from randomized and quasi-randomized controlled trials of thyroid hormone therapy in preterm infants. Thyroid hormones have effects not only on neurological development but also on the respiratory and cardiovascular systems, and somatic growth. Therefore, studies that examined either neonatal morbidity, mortality and/or long term neurodevelopment were included in this review.

Objectives

To assess whether thyroid hormone therapy in preterm infants without congenital hypothyroidism results in clinically important changes in neonatal and long term outcomes in terms of both benefits and harms. To determine the evidence for the use of different thyroid hormone preparations, doses and timing of treatment by subgroup analysis of the trials.

Criteria for considering studies for this review

Types of studies

Trials using random or quasi-random patient allocation to treatment or control.

Types of participants

Studies that enrolled and treated preterm infants in the neonatal period.

Types of interventions

Thyroid hormone therapy (either treatment or prophylaxis) compared to control (placebo or no therapy). Thyroid hormone therapy could be either thyroxine, triiodothyronine or both.

Types of outcome measures

Primary clinical outcome measures were mortality (either neonatal or prior to discharge) and neurodevelopmental status at follow up. Neurodevelopmental outcome was categorized as:

  1. Abnormal mental developmental (a development or intelligence quotient > 2 standard deviations below the mean of a standardized test),
  2. Abnormal neurological outcome (infants with either abnormal mental development or definite cerebral palsy or hearing deficit requiring aids or visual acuity < 6/60),
  3. Motor deficits, and
  4. Sensorineural impairments.

Secondary outcome measures included important neonatal morbidities and measures of potential adverse effects of thyroid hormone treatment:

  1. Severity of respiratory disease (use of IPPV or CPAP, maximum ventilation requirements, use of rescue therapies, incidence of air leak, duration of mechanical ventilation),
  2. Patent ductus arteriosus diagnosed clinically or by echocardiography,
  3. Intraventricular hemorrhage diagnosed by ultrasound or postmortem
  4. Periventricular leucomalacia diagnosed by ultrasound or postmortem,
  5. Chronic lung disease defined as oxygen at 28 days post-natal age or 36 weeks post-menstrual age,
  6. Necrotizing enterocolitis,
  7. Retinopathy of prematurity, and
  8. Growth.

Search strategy for identification of studies

The standard search strategy of the Neonatal Review Group was used. This included searches of the Oxford Database of Perinatal Trials, Cochrane Controlled Trials Register, MEDLINE, PREMEDLINE, EMBASE, previous reviews including cross references, abstracts, conferences (SPR-PAS 1998-2001and PSANZ 1998-2001), expert informants (authors of published trials and neonatologists) and journal handsearching in the English language.

MEDLINE (1966 to May 2001) was searched using MeSH terms '(infant-newborn or infant-premature) and (thyroxine or triiodothyronine)' and text words using '(hypothyroxinemia or hypothyroxinaemia or thyroxine or triiodothyronine) and [MeSH terms] (infant-newborn or infant-premature)'. EMBASE (1980 to 2001 Week 22) and PREMEDLINE (June 2001) were searched using terms '(hypothyroxinemia or hypothyroxinaemia or thyroxine or triiodothyronine) and newborn. The Oxford Database of Perinatal Trials was searched using the term 'thyroid disease', and the Cochrane Controlled Trials Register (2001, Issue 2) was searched using 'thyroxine or triiodothyronine or hypothyroxinemia or hypothyroxinaemia'.

[top]

Methods of the review

Assessment of trial quality, data extraction and synthesis of data, using relative risk (RR) and weighted mean difference (WMD), were performed using standard methods of the Cochrane Collaboration and its Neonatal Review Group. The study by van Wassenaer 1997 was the subject of several publications. Additional data was obtained from the authors of three studies (Chowdhry 1984, Vanhole 1997 and van Wassenaer 1997). Data were synthesized using relative risk (RR) and risk difference (RD). From 1/RD the number needed to treat (NNT) for benefits and the number needed to harm (NNH) for adverse effects was calculated. All results are given with the 95% confidence intervals unless otherwise stated. Attempts to contact authors of the THORN study have been unsuccessful.

Trials were included in the review if there was > 80% ascertainment of the outcome in survivors analyzed on an intention to treat basis.

Trials using thyroid hormone therapy in unselected preterm infants, infants with respiratory distress and infants with low thyroid hormone levels were included in the review. This allows the review to explore differences in treatment effect in infants with different levels of risk for abnormal outcomes.

Subgroup analysis was performed for those trials using good methodology as defined by the use of a random method of allocation to treatment or control, if steps were taken to ensure allocation concealment and if there was at least 90% follow up of survivors.

Subgroup analyses were performed according to thyroid hormone therapy used (triiodothyronine or thyroxine); timing of thyroid hormone therapy (early < 24 hours age); dosage of thyroid hormone preparation used; use of thyroid hormones as prophylaxis (of all preterm infants irrespective of thyroid hormone status); and as treatment (of preterm infants with low thyroid hormone levels).

Subgroup analysis by gestational age at birth performed by one study (van Wassenaer 1997) suggesting a difference in effect at different gestations was not prespecified.

Only trials that commenced thyroid hormone treatment early (< 24 hours) were included in the analyses of short term respiratory outcomes.

Where appropriate, data for some outcomes were pooled despite differences between studies in the method of ascertainment of the outcome (see description of studies for definitions used in individual studies). These outcomes included chronic lung disease (oxygen at 28 days), patent ductus arteriosus and intraventricular hemorrhage (any grade and grade 3 or 4).

Description of studies

Studies included in this review are: Chowdhry 1984, Amato 1989, van Wassenaer 1997, Vanhole 1997 and Smith 2000. Details of each study are given in the table 'Characteristics of Included Studies' and references. No ongoing trials of thyroid hormone replacement are currently known. A trial of thyroid hormone treatment in preterm infants has been proposed and is in the design phase (Golombek).

Studies excluded from this review are: Schonberger 1981 (used alternation and included non-alternated infants in control group), Eggermont 1984 (non-randomized trial), Amato 1988 (used alternation and excluded 25% of infants after alternation), and van Wassenaer 1993 (non-randomised study).

Studies awaiting assessment: THORN trial (Thyroid hormone replacement in neonates).

Types of participants:
All studies enrolled very preterm infants but used different entry criteria. Chowdhry 1984 enrolled infants < 1250g and 25-28 weeks' gestation who were demonstrated to have hypothyroxinemia (thyroxine < 4 micrograms/dL and thyrotropin < 20 IU/L) on two occasions. Thyroid function was not an entry criterion for any other study. Amato 1989 enrolled infants born < 32 weeks' gestation in > 40% oxygen with a clinical and x-ray diagnosis of respiratory distress syndrome. Van Wassenaer 1997 enrolled infants born 25-29 weeks' gestation. Vanhole 1997 enrolled infants born 25-30 weeks' gestation with available cord blood for baseline thyroid function. Smith 2000 enrolled infants < 32 weeks' gestation and 600-1500g. Three studies (van Wassenaer 1997, Vanhole 1997 and Smith 2000) excluded infants with congenital malformations and maternal endocrine disease.

Interventions:
Three studies (Chowdhry 1984, van Wassenaer 1997, Vanhole 1997 and Smith 2000) compared thyroxine treatment to placebo, whereas Amato 1989 compared triiodothyronine treatment to no treatment. Chowdhry 1984 treated infants with thyroxine 10 micrograms/kg/day given intramuscularly until tolerating feeds, then orally. Amato 1989 treated infants with L-triiodothyronine 50 micrograms/day in 2 divided doses for 2 days intravenously starting on the first day. Van Wassenaer 1997 treated infants with thyroxine 8 micrograms/kg birthweight (intravenously until tolerating feeds, then orally), from 12-24 hours of age, for 6 weeks. Vanhole 1997 treated infants with thyroxine 20 micrograms/kg/day intravenously for 2 weeks from the first day. Smith 2000 treated infants, starting before 48 hours age, with thyroxine 10 micrograms/kg/day intravenously till tolerating feeds, then 20 micrograms/kg/day for a total 21 days.

Outcomes:
A principal outcome was stated by two studies, van Wassenaer 1997 (Bayley MDI at 24 months) and Smith 2000 (oxygen at 28 days). Other studies stated broader objectives: Amato 1989 assessed the effect of thyroxine on the course and severity of respiratory distress syndrome. Chowdhry 1984 assessed the effect of thyroid hormone therapy on growth and development. Vanhole 1997 assessed its impact on a variety of endocrine and clinical outcomes.

Mortality was assessed by all the included studies with data for neonatal mortality available for Chowdhry 1984, van Wassenaer 1997, Vanhole 1997 and Smith 2000, and mortality to time of discharge for all studies. Amato 1989 did not document time of death of four infants (all died from intraventricular hemorrhage). These infants were included in deaths to discharge.

Neurodevelopment was assessed by van Wassenaer 1997 and Vanhole 1997. The Bayley MDI and the Bayley Psychomotor Development Index (PDI) were used by van Wassenaer 1997 at 12 and 24 months corrected age and Vanhole 1997 at 7 months corrected age. van Wassenaer 1997 assessed neurological outcome at 6 and 12 months (method of Touwen) and 24 months (method of Hempel). At mean age 5.7 years, van Wassenaer 1997 assessed cognitive (RAKIT IQ score), behaviour, motor and neurologic (method of Touwen) outcome. As only 38% of survivors received neurodevelopmental assessment in the study by Chowdhry 1984, this study is excluded from any analysis of neurodevelopmental outcome.

Data for chronic lung disease defined as oxygen requirements at 28 days were obtained from the studies by Chowdhry 1984, Amato 1989, van Wassenaer 1997 and Smith 2000. Vanhole 1997 defined chronic lung disease as oxygen requirements at 28 days with an abnormal chest x-ray. These data are pooled. Data were available from Chowdhry 1984, van Wassenaer 1997 and Vanhole 1997 for infants with chronic lung disease as defined by oxygen requirements at 36 weeks post-menstrual age.

Patent ductus was diagnosed clinically by Amato 1989 and clinically with ultrasound confirmation by van Wassenaer 1997. Intraventricular hemorrhage was detected by ultrasound by Amato 1989, and ultrasound and post-mortem by van Wassenaer 1997. Data were obtained from the author for intraventricular hemorrhage and patent ductus arteriosus from the study by Vanhole 1997 and for patent ductus arteriosus from the study by Chowdhry 1984, although no definitions were given.

Methodological quality of included studies

Randomization: four studies used a random method of allocation to treatment. Chowdhry 1984 randomized infants in pharmacy by "drawing cards". van Wassenaer 1997 used computerised block randomization. Vanhole 1997 used randomized envelopes. Smith 2000 used sequentially numbered envelopes stratified by center. Amato 1989 was a "quasi-random" study, treating alternate patients.

There were no significant differences between treatment and control groups after randomization in terms of birth weights, gestational ages, Apgar scores or the male/female ratio in any of the studies. van Wassenaer 1997 had fewer infants < 27 weeks gestation (19 versus 27) and more very small for gestational age infants (5 versus 2) in the treatment group compared to the control group. In the neurodevelopmental follow up of 25-26 weeks' gestation survivors, van Wassenaer 1997 found a significant lower mean birth weight in the treatment group (n = 13) compared to the placebo group (n = 18) (856 ± 153g versus 914 ± 111g; p = 0.04). Smith 2000 had fewer infants randomised to control (20) than to treatment (29). No reason was given (reply from the author awaited).

Blinding of treatment: Chowdhry 1984, van Wassenaer 1997, Vanhole 1997 and Smith 2000 used placebo treatment in the control arm. Amato 1989 was unblinded by the use of alternation and no treatment in the control arm.

Blinding of outcome assessment: Chowdhry 1984, van Wassenaer 1997, Vanhole 1997 and Smith 2000 performed blinded assessment of clinical and neurodevelopmental outcomes. Amato 1989 did not state whether the assessment of clinical outcomes was blinded.

Exclusions after randomization: Chowdhry 1984 followed all infants up for short term clinical outcomes but assessed only 38% of survivors for neurodevelopmental. Amato 1989 excluded 6 (12%) infants (3 treatment and 3 controls) after alternation as they had respiratory diagnoses other than respiratory distress syndrome. Van Wassenaer 1997 had 7 (4%) losses to neurodevelopmental follow up at 24 months corrected age (4 treatment and 3 controls) due to withdrawals from trial or the diagnosis of a congenital anomaly, and 9 (5 treatment and 4 controls) at 5.7 years. Vanhole 1997 excluded 6 (15%) neonatal deaths from analysis of neonatal morbidities, but assessed neurodevelopment of all survivors at 7 months corrected age. Smith 2000 excluded 2 infants (4%) from the control group.

Studies with good methodology: four studies (Chowdhry 1984, van Wassenaer 1997, Vanhole 1997, Smith 2000) were of good methodology with adequate randomization, allocation concealment and > 90% follow up of survivors. Amato 1989 was excluded due to use of non-random allocation to treatment (alternation) and 12% post alternation exclusions, and Chowdhry 1984 was excluded from analysis of neurodevelopmental outcomes due to inadequate follow up. van Wassenaer 1997, Vanhole 1997 and Smith 2000 were included for all available outcomes.

[top]

Results

See "tables of analyses".

Neurodevelopment:

No significant difference was found in any measure of neurodevelopmental outcome in infants who received neonatal thyroid hormone replacement compared to controls. There was no significant difference in the mortality rate or combined mortality and abnormal neurological outcome in infants who received thyroid hormone treatment compared to controls.

Neurodevelopment: van Wassenaer 1997 found no difference in the risk of abnormal mental development (> 2sd below mean) at 6, 12, 24 months or 5.7 years (at 5.7 years relative risk 0.66, 95% CI 0.22, 1.99) in infants receiving thyroid hormones compared to controls. There was no significant difference in the risk of cerebral palsy assessed at 5.7 years (RR 0.72, 95% CI 0.28, 1.84). Meta-analysis of two studies, van Wassenaer 1997 and Vanhole 1997, found no difference in the Bayley MDI (weighted mean difference -1.14, 95% CI -5.46, 3.19) and PDI (WMD 0.22 points, 95% CI -4.80, 5.24) at 7-12 months. van Wassenaer 1997 found no difference in the Bayley MDI (MD -3.50 points, 95% CI -11.21, 4.21) and PDI (MD 3.10 points, 95% CI -3.31, 9.51) at 24 months. van Wassenaer 1997 found similar RAKIT IQ scores at 5.7 years (MD -2.10 points, 95% CI -7.91, 3.71). At 2 years the risk of a clinically significant Child Behavioural Checklist (CBCL) score (RR 0.81, 95% CI 0.35, 1.86) was not significantly different. At 5.7 years, the risk of a clinically significant CBCL score (RR 1.17, 95% CI 0.57, 2.40) and Teacher Report Form (TRF) score (RR 1.08, 95% CI 0.55, 2.10) were not significantly different.

In post hoc subgroup analyses, van Wassenaer 1997 found a significant reduction in the incidence of abnormal mental development at 24 months in infants born 25-26 weeks gestation (RR 0.10, 95% CI 0.01, 0.70). However, there was no significant difference in abnormal mental development in infants born 27-29 weeks' gestation (RR 1.32, 95% CI 0.46, 3.82). There was a significant increase in the Bayley MDI at 24 months in treated infants of 25-26 weeks' gestation (MD 18.0 points, 95% CI 0.3 to 35.7), but a significant decrease in the Bayley MDI at 24 months in treated infants of 27-29 weeks' gestation (MD -10.0 points, 95% CI -18.4 to -1.6). At 5.7 years, the difference in the RAKIT IQ score was no longer significant in infants 25-26 weeks (MD 9.50 points, 95% CI -0.88, 19.88) but was significant in infants 29 weeks' gestation (MD -14.6 points, 95% CI -25.14, -4.06). There were no differences at any gestation in the psychomotor scores or incidences of cerebral palsy. Infants 25-27 weeks had an increased incidence of clinically significant CBCL problem scores.

Death:

Meta-analysis of all five trials does not support a decreased incidence of mortality to discharge in infants receiving thyroid hormone treatment (typical RR 0.70, 95% CI 0.42 to 1.17).

Death or cerebral palsy: data were available from one study, van Wassenaer 1997 with no significant difference at 5.7 years (RR 0.70, 95% CI 0.43 to 1.14).

Neonatal morbidity:

Severity of respiratory disease:

Ventilation and oxygen requirements: no data could be pooled for meta-analysis. No data were obtained for short term outcomes such as maximal oxygenation index or use of rescue therapies. Amato 1989 found a significant reduction at 12 to 72 hours in the fraction of inspired oxygen, but not peak inspiratory pressure, mean airway pressure or oxygenation, in the triiodothyronine treated infants. Vanhole 1997 found no difference in the inspired oxygen requirements in the first 6 days of life.

Airleak: one study, Amato 1989, found no difference in pulmonary airleak in infants with respiratory distress syndrome treated with thyroid hormones (RR 0.75, 95% CI 0.19 to 2.97).

Duration of mechanical ventilation: meta-analysis of three trials (Amato 1989, Vanhole 1997, Smith 2000) found no significant difference in duration of mechanical ventilation (WMD -2.44 days, 95% CI -5.41 to 0.53). Chowdhry 1984 reported no difference in the duration of ventilation (no data given).

Chronic lung disease: meta-analysis of all five trials found no significant difference in incidence of chronic lung disease at 28 days postnatal age (typical RR 0.96, 95% CI 0.75, 1.23). Meta-analysis of three trials (Chowdhry 1984, van Wassenaer 1997, Vanhole 1997) reported no significant difference in incidence of oxygen requirements at 36 weeks postmenstrual age (typical RR 1.02, 95% CI 0.64, 1.64).

Patent ductus arteriosus: meta-analysis of five trials found no significant difference in incidence of clinically or echocardiographically diagnosed patent ductus arteriosus (typical RR 0.76, 95% CI 0.56, 1.03).

Intraventricular haemorrhage: meta-analysis of three trials (Amato 1989, van Wassenaer 1997, Vanhole 1997) does not support any benefit of thyroid hormone treatment on the incidence of intraventricular hemorrhage (typical RR 1.20, 95% CI 0.88, 1.63). Meta-analysis of three trials (van Wassenaer 1997, Vanhole 1997, Smith 2000) found no difference in incidence of severe (grade 3 or 4) intraventricular hemorrhage (typical RR 1.07, 95% CI 0.55, 2.11).

Periventricular leucomalacia: meta-analysis of two trials (van Wassenaer 1997, Smith 2000) found no significant difference in incidence of periventricular leucomalacia (typical RR 0.84, 95% CI 0.16, 4.38).

Necrotizing enterocolitis: meta-analysis of two trials (Chowdhry 1984, Amato 1989) found no significant difference in incidence of necrotizing enterocolitis (typical RR 0.35, 95% CI 0.04, 3.16).

Retinopathy of prematurity: meta-analysis of three trials (Chowdhry 1984, Vanhole 1997, van Wassenaer 1997) found no significant difference in incidence of retinopathy of prematurity (typical RR 0.93, 95% CI 0.49, 1.78). Amato 1989 reported only cicatricial 'retrolental fibroplasia' (RR 0.50, 95% CI 0.05, 5.08).

Growth: one study reported data on rates of growth (Chowdhry 1984). There was no significant difference in the rates of weight increase (MD 0.5 grams/day, 95% CI -2.5, 3.6), rates of growth in head circumference (MD 0.02 cm/week, 95% CI -0.15, 0.19), or growth in length (MD -0.06 cm/week, 95% CI -0.24, 0.12). Vanhole 1997 also found no difference in growth rates of infants receiving thyroxine treatment (displayed as graph in paper).

Other outcomes: in a subgroup of infants, van Wassenaer 1997 found no effect of thyroxine therapy on somatosensory evoked potentials at 2 weeks, term and 6 months of age (reference d). At the same intervals, motor nerve conduction velocity was also not significantly affected in 80 treated infants compared to 83 controls (reference e). Vanhole 1997 found no effect of thyroxine treatment on heart rate, percentage of enteral fluid intake, or on weight gain. One study (Smith 2000) reported an reduction in sepsis in infants receiving thyroxine (RR 0.51, 95% CI 0.26, 0.98) which was reported in the article as not significant using the chi-square test at the p = 0.05 level.

Studies with good methodology: the four studies of good methodology (Chowdhry 1984, van Wassenaer 1997, Vanhole 1997, Smith 2000) all used thyroxine in the treatment group. Chowdhry 1984 was excluded from analysis of neurodevelopmental outcomes. The results of this review are not changed by incorporating only studies with good methodology. The neurodevelopmental outcomes are the same, and meta-analysis of the four studies shows no difference in neonatal mortality. All other neonatal outcomes are similar.

Thyroid hormone preparation used: four studies (Chowdhry 1984, van Wassenaer 1997, Vanhole 1997, Smith 2000) used thyroxine. The studies and results are the same as those described above in "studies with good methodology". Only Amato 1989 used triiodothyronine (for 2 days) to determine the effect of therapy on the severity of respiratory distress syndrome. Neurodevelopment was not assessed. No significant difference in mortality was found (RR 1.00, 95% CI 0.15 to 6.48). Apart from a reduced fraction of inspired oxygen in the first 72 hours, no significant differences were found in the incidences of any other neonatal outcome including air leak, duration of mechanical ventilation, oxygen at 28 days, patent ductus arteriosus, intraventricular hemorrhage, necrotizing enterocolitis or cicatricial 'retrolental fibroplasia'.

Chowdhry 1984 was the only study to use thyroid hormones as treatment for infants with transiently low thyroid hormone levels (transient neonatal hypothyroxinemia). Neurodevelopmental follow-up was inadequate. No significant difference in mortality to discharge was found (RR 1.09, 95% CI 0.08, 15.42).

Thyroid hormones were used as prophylaxis (in unselected preterm infants without documented low thyroid hormone levels) by four studies (Amato 1989, van Wassenaer 1997, Vanhole 1997, Smith 2000). Three of these studies (Amato 1989; Vanhole 1997; van Wassenaer 1997) commenced treatment within 24 hours of birth, with Smith 2000 commencing treatment before 48 hours. Amato 1989 was the only study to find a difference in a short term measure of respiratory function with a lower inspired oxygen in the treatment group in the first 72 hours. No other differences were found in any other respiratory outcome in any of the studies including incidence of airleak, duration of mechanical ventilation or incidence of chronic lung disease (at 28 days or 36 weeks).

Various doses of thyroxine were used in three trials with Chowdhry 1984 using 10 micrograms/kg/day from day 15 of life for 7 weeks, van Wassenaer 1997 using 8 micrograms/day from day one of life for 6 weeks, Vanhole 1997 using 20 micrograms/kg/day from day 1 for 2 weeks and Smith 2000 10 micrograms/kg/day IV then 20 micrograms/kg/day orally for total 21 days. Neurodevelopmental outcomes are only available for van Wassenaer 1997 and Vanhole 1997 who both used thyroxine from day one. Van Wassenaer 1997 found no significant difference in incidence of cerebral palsy or mental development at 5.7 years. Vanhole 1997 found a nonsignificant reduction in the Bayley MDI at 7 months corrected age. In a post hoc subgroup analysis, van Wassenaer 1997 found a difference in effect according to gestation, with infants born at 25-26 weeks gestation having significantly higher mean mental scores in the treatment group, but infants born at 27-29 weeks gestation having significantly lower scores in the treatment group. The differences in the cognitive scores for infants 25-26 weeks were no longer significant at 5.7 years but were for infants at 29 weeks. No other differences were found for other outcomes including mortality, respiratory distress syndrome, and chronic lung disease.

Discussion

Nine studies were identified that compared thyroid hormone treatment to placebo or no treatment. Four studies were excluded: Schonberger 1981 (used alternation and included non-alternated infants in control group), Eggermont 1984 (non-randomized trial), Amato 1988 (used alternation and excluded 25% of infants after alternation), and van Wassenaer 1993 (non-randomised study).

Five studies were included in this review (Chowdhry 1984, Amato 1989, van Wassenaer 1997, Vanhole 1997, Smith 2000). All studies enrolled preterm infants but used different entry criteria. Amato 1989 studied infants < 32 weeks' gestation with respiratory distress syndrome. Chowdhry 1984 studied infants 25-28 weeks gestation and < 1250 grams with low thyroxine levels (T4 < 4 microgram/dl) on two occasions. van Wassenaer 1997 studied infants 25-29 weeks gestation, Vanhole 1997 infants 25-30 weeks gestation with cord blood available and Smith 2000 infants < 32 weeks.

Limitations to this review:

Major limitations of this review include: the small number of infants enrolled in trials incorporated in the review resulting in limited power of this review to detect moderate but clinically important differences in effect of thyroid hormone therapy in preterm infants; and the fact that studies used different treatments, doses and enrolment criteria limiting the ability to make conclusions from pooled results. This review examines the clinical effect of thyroid hormone therapy and does not address the question of what levels of thyroid hormones should be targeted in preterm infants, and what types and doses of thyroid hormones should be used to achieve these levels.

For primary outcomes, limited comparable data were available for meta-analytic pooling of results for neurodevelopment, but data were complete for all studies with respect to mortality to discharge.

Limitations of studies:

Two studies (van Wassenaer 1997 and Vanhole 1997) with adequate neurodevelopmental follow up were of good methodology, whereas one study (Chowdhry 1984) with good methodology for short term outcomes had inadequate neurodevelopmental follow up. Amato 1989 used alternation and had > 10% exclusions after alternation. The largest study with 200 infants, van Wassenaer 1997, had non-significant differences in neonatal risk factors in the two groups after randomization, with fewer infants < 27 weeks and more infants < 3rd percentile for weight in the treatment group. Smith 2000 had more infants randomised to the treatment group (29 versus 20).

Four studies used thyroid hormones prophylactically in infants at risk of transient hypothyroxinemia, whereas only one study (Chowdhry 1984) treated only infants with demonstrated hypothyroxinemia. This study did not achieve adequate neurodevelopmental follow up. No study with adequate neurodevelopmental follow up treated infants with demonstrated hypothyroxinemia or was restricted to those most likely to benefit (infants < 27 weeks' gestation).

The studies were small (four enrolling between 23 and 49 infants), with van Wassenaer 1997 the largest enrolling a total of 200 infants. The small numbers of patients studied limits the power of this review to detect a moderate, but potentially important change in neurodevelopmental outcome.

The review does not support any beneficial effect of thyroid hormone treatment in preterm infants on subsequent neurodevelopment. van Wassenaer 1997 found no difference in the risk of cerebral palsy or abnormal mental development at 6, 12, 24 months or 5.7 years. Both van Wassenaer 1997 and Vanhole 1997 found no significant differences in developmental or intelligence quotients in treated infants compared with controls.

In post hoc subgroup analysis, van Wassenaer 1997 found that there was a significant increase in the Bayley MDI at 24 months in treated infants of 25-26 weeks gestation, but a significant decrease in the Bayley MDI at 24 months in treated infants of 27-29 weeks' gestation. This finding in infants of 24-25 weeks' gestation should be treated with caution as the analysis was not prespecified, there were imbalances in risk factors between the groups after randomization, the difference in IQ scores for infants 25-26 weeks gestation was no longer significant at 5.7 years and the overall finding of no effect for the combined gestations. This does not preclude a benefit of thyroid hormone therapy in preterm infants < 27 weeks' gestation.

Meta-analysis of all five trials does not support a decrease in mortality in infants receiving thyroid hormone treatment. The trend to a reduction in combined mortality and cerebral palsy in one study (van Wassenaer 1997) does not reach significance. If the effect on combined mortality and cerebral palsy is of a magnitude found by van Wassenaer 1997 (RR 0.70, 95% CI 0.43, 1.14), then this would be a clinically important effect.

For secondary outcomes, no significant differences were found for any neonatal morbidity including chronic lung disease, patent ductus arteriosus, intraventricular hemorrhage, retinopathy of prematurity and growth. Fraction of inspired oxygen was lower in infants receiving triiodothyronine in one small quasi-randomized study, but not in infants receiving thyroxine in a randomized study (Vanhole 1997). The reduction in sepsis (Smith 2000) was reported as non-significant by the authors. No other study reported this outcome. No other significant differences were found to suggest a reduced severity of respiratory distress syndrome including incidence of airleak, duration of mechanical ventilation or incidence of chronic lung disease.

There is no apparent heterogeneity in the findings of the different studies. The conclusions are not altered by excluding studies with inadequate methodology. The review finds no evidence of benefit from any individual thyroid hormone preparation, dose, timing or duration of treatment used by the included studies.

Reviewers' conclusions

Implications for practice

This review does not support the use of thyroid hormones in preterm infants to reduce neonatal mortality, improve neurodevelopmental outcome or reduce the severity of respiratory distress syndrome. The post hoc analyses of data from one study (van Wassenaer 1997) which showed benefits in infants 24-25 weeks' gestation should be treated with caution. There is no evidence of benefit from treatment of transient hypothyroxinemia in preterm infants.

Implications for research

The small number of infants included in trials incorporated in this review limits the power of the meta-analysis to detect clinically important differences in neonatal outcomes. The findings of this review do not preclude a benefit of thyroid hormone treatment in extremely preterm infants or infants with hypothyroxinemia.

Future trials should be of sufficient size to detect clinically important differences in neurodevelopmental outcomes. To detect the magnitude of effect on the incidence of abnormal neurological outcome as found in the study by van Wassenaer 1997, a trial would need to enrol 1670 infants. This number would provide 80% power to detect a 40% reduction in death or cerebral palsy (from 12 to 7%) at the 95% confidence level (assuming a 25% mortality rate).

Future trials should consider enrolling only those infants most likely to benefit from thyroid hormone treatment. This may include enrolling only infants < 27 weeks' gestation. Trials should incorporate neurodevelopmental follow up most likely to detect abnormalities resulting from low thyroid hormone levels in preterm infants. This includes the use of standardized tests of mental and psychomotor development, examination for cerebral palsy, as well as follow up to school age to detect cognitive deficits and evidence of school failure.

Acknowledgements

Dr Parveen Chowdhry, Professor F de Zegher, Dr Aleid van Wassenaer and Dr Christine Vanhole for kindly responding to the request for additional information. Professor David Henderson-Smart assisted with methodological issues with this review.

Potential conflict of interest

  • None noted.

[top]

Characteristics of studies

Characteristics of Included Studies

Study Methods Participants Interventions Outcomes Notes Allocation concealment
Amato 1989 Quasi-random study: alternate assignment.
Single center.
Blinding of randomization: no.
Blinding of intervention: not stated.
Complete follow up: no, 6 (12%) post-randomisation exclusions (3 treatment and 3 controls).
Blinding of outcome measurement: not stated.
Power calculation performed: not stated.
Inclusion criteria:
Fifty preterm infants < 32 weeks gestation with respiratory distress syndrome diagnosed by presence of respiratory distress and chest x-ray changes.
FiO2 > 0.4
No antenatal steroids.
Exclusion criteria:
Not respiratory distress syndrome.
Treatment group: mean birth weight (± sd) 1180 ± 430g, mean gestation 30.1 ± 1.9 weeks (n = 25).
Controls: mean birth weight 1210 ± 380g, mean gestation 29.2 ± 2.1 weeks (n = 25).
Treatment: L-triiodothyronine 50 micrograms/day in 2 divided doses for 2 days IV, commenced day 1.
Control: no treatment.
Mortality to discharge.
Peak oxygen concentrations.
Duration mechanical ventilation.
Development of major complications of RDS: periventricular hemorrrhage, air leak, chronic lung disease (O2 at 28 days), patent ductus arteriosus, necrotizing enterocolitis, retinopathy of prematurity (cicatricial disease only).
C
Chowdhry 1984 Random study: yes - randomized in pharmacy by "drawing cards".
Single center.
Blinding of randomization: yes.
Blinding of intervention: yes, placebo controlled.
Complete follow up: no, 15 (65%) lost to neurodevelopmental follow up. All infants had mortality, thyroid function and growth assessed.
Blinding of outcome measurement: yes.
Power calculation performed: not stated.
Twenty three preterm infants < 1250g and 25-28 weeks' gestation with T4 less than/or equal to 4 microg/dl and TSH less than/or equal to 20 IU/L on two occasions.
Treated group: mean birth weight (± sd) 834 ± 182g, mean gestation 26.5 ± 10.6 weeks (n = 11).
Control group: mean birth weight 738 ± 162g, mean gestation 26.3 ± 3.4 weeks (n = 12).
Treatment: from day 15, approximately 7 weeks treatment with thyroxine 10 microg/kg/day IM (orally when tolerating feeds). Increased to 15 micrograms/kg/day if no increase in serum T4 after 1 week.
Controls: placebo.
Neurodevelopment at 12 months: Bayley Mental Development Index and Bayley Psychomotor Index.
Neonatal mortality.
Mean time to T4 > 5 microg/dl.
Weight gain (mean daily weight gain = g/day).
Growth of head circumference = cm/week.
Growth of length = cm/week.
Five treated and 3 controls evaluated at 12 months.
Two treated and 2 control infants evaluated at 24 months.
A
Smith 2000 Random study: yes, sequentially numbered envelopes in blocks of 8 allocated to individual centres.
Two centers.
Blinding of intervention: yes, placebo controlled.
Complete follow up: no, 2 losses in control (> 32 weeks, thyroid disease).
Blinding of outcome measurement: probably, but one investigator examined thyroid tests to determine need for thyroid treatment.
Power calculation performed: yes, 300 needed to detect reduced incidence of oxygen at 28 days from 30 to 15%.
Forty nine infants.
Inclusion criteria: Infants < 32 weeks, birthweight 600-1500g.
Exclusion criteria: maternal thyroid disease or illicit drug use, infant lethal malformation or congenital thyroid disease.
Second of twins not enrolled.
Treatment group: (± sd) 28.6 (0.4) weeks gestation, 1074g (46) birthweight (n = 29).
Control group: 28.8 (0.5) weeks, 1037g (66) birthweight (n = 18, 2 losses).
Treatment: IV thyroxine 10 micrograms/kg/day started before 48 hours age. Oral thyroxine 20 microgram/kg/day commenced as soon as oral feedings tolerated.
Control: placebo (saline).
Continued for 21 days.
Primary outcome: oxygen at 28 days.
Secondary: death before 12 months, intraventricular hemorrhage, periventricular leucomalacia, patent ductus arteriosus, sepsis and necrotizing enterocolitis.
Study stopped early due to minimal effect seen on oxygen at 28 days. A
van Wassenaer 1997 Random study: computerised block (size 10) randomization. Not stratified.
Single center.
Blinding of randomization: yes.
Blinding of intervention: yes, placebo controlled.
Complete follow up: yes: 96% of survivors had neurodevelopmental followup. Treatment group had 4 losses and control group 3 (congenital abnormality or withdrawn). At 2.7 years, losses were treatment 5 and control 4.
Blinding of outcome measurement: yes.
Power calculation performed: not stated.
Two hundred preterm infants.
Inclusion criteria: preterm infants 25-29 weeks.
Exclusion criteria: severe congenital malformations, maternal endocrine disease or illicit drug use.
Treatment group: mean birth weight (± sd) 1078 ± 218g, mean gestation 197 ± 8 days (n = 100).
Controls: mean birth weight 1077 ± 239g, mean gestation 196 ± 9 days (n = 100).
Treatment: thyroxine 8 micrograms/kg daily (IV until tolerating feeds, then oral) for 6 weeks, starting 12-24 hours after birth.
Control: placebo for 6 weeks.
Neurodevelopment at 6 and 12 months (method of Touwen).
Neurodevelopment at 24 months (method of Hempel).
Bayley Mental Development Index at 6, 12 and 24 months.
Bayley Psychomotor Index at 6, 12 and 24 months corrected age. Child Behaviour Checklist (CBCL) for ages 2-3.
At mean age 5.7 years: Revised Amsterdam Children's Intelligence Test (RAKIT), CBCL and Teacher Report Form (TRF), Movement Assessment Battery for Children and neurologic outcome using method of Touwen (minor neurological dysfunction [MND] and cerebral palsy.
Mortality (to time of discharge), respiratory distress syndrome, patent ductus arteriosus, intraventricular hemorrhage,
cerebral ischemic lesions, chronic lung disease (at 36 weeks), retinopathy of prematurity.
Subgroup analysis stratified for gestation (24-25 weeks and 26-29 weeks) not prespecified. A
Vanhole 1997 Random study: used sealed envelopes.
Single center.
Blinding of randomization: yes.
Blinding of intervention: yes, placebo controlled.
Complete follow up: yes - all survivors had neurodevelopmental follow up. 6 (15%) neonatal deaths excuded from analysis of neonatal morbidities (CLD and RDS).
Blinding of outcome measurement: yes.
Power calculation performed: not stated.
Forty preterm infants:
Inclusion criteria: inborn, gestation 25-30 weeks, cord blood available.
Exclusion criteria: congenital malformations and maternal thyroid disease.
Treatment group: mean birth weight (± sd) 1191 ± 299g, mean gestation 28.4 ± 1.7 weeks (n = 20).
Control group: mean birth weight 1206 ± 411g, mean gestation 28.5 ± 1.9 weeks (n = 20).
The mothers of 11/17 of the treatment group and 13/17 of controls received antenatal steroids and TRH.

Treatment: thyroxine 20 microgram/kg/day IV for two weeks from day 1.

Control: placebo.

Mean Bayley Mental Development Index at 7 months corrected age
Mean Bayley Psychomotor Index at 7 months corrected age
Neonatal mortality
Respiratory distress syndrome
Chronic lung disease (oxygen at 28 days and CXR changes)
Heart rate, daily inspired oxygen, weekly weight gain, and enteral fluid intake
Also evaluated thyroid hormone (T4, T3 and reverse T3), prolactin and growth hormone levels in infants and effects of dopamine infusions. A

Characteristics of excluded studies

Study Reason for exclusion
Amato 1988 Inadequate randomisation performed. Used alternate assignment and excluded 25% after trial entry as they did not have respiratory distress syndrome.
Eggermont 1984 Non-randomized cohort comparison. Thyroxine given to 'sick' preterm infants and compared to a control group of 'non-sick' preterm infants.
Schonberger 1981 Inadequate randomisation. Used alternation and included 5 infants who were not alternated as controls.
van Wassenaer 1993 Not a randomised study. Used historical controls to compare the effect of three different doses of thyroxine on neonatal thyroid hormone levels. Dose was varied over 3 consecutive time periods (6, 8 and 10 micrograms/kg/day).

Characteristics of ongoing studies

Study Trial name or title Participants Interventions Outcomes Starting date Contact information Notes
Golombek Hypothyroxemia trial Very preterm infants (< 28 weeks) Thyroid hormones - type and dose to be determined Include neurodevelopment To be announced SERGIO_GOLOMBECK@NYMC.EDU

[top]

References to studies

Included studies

Amato 1989

{published data only}

Amato M, Guggisberg C, Schneider H. Postnatal triiodothyronine replacement and respiratory distress syndrome of the preterm infant. Horm Res 1989;32:213-7.

Chowdhry 1984

{published and unpublished data}

Chowdhry P, Scanlon JW, Auerbach R, Abbassi V. Results of controlled double-blind study of thyroid replacement in very low-birth-weight premature infants with hypothyroxinemia. Pediatrics 1984;73:301-5.

Smith 2000

{published data only}

Smith LM, Leake RD, Berman N, Villanueva S, Brasel JA. Postnatal thyroxine supplementation in infants less than 32 weeks' gestation: effects on pulmonary morbidity. J Perinatol 2000;20:427-31.

van Wassenaer 1997

{published and unpublished data}

* van Wassenaer AG, Kok JH, de Vijlder JJ, Briet JM, Smit BJ, Tamminga P, van Baar A, Dekker FW, Vulsma T. Effects of thyroxine supplementation on neurologic development in infants born at less than 30 weeks' gestation. N Engl J Med 1997;336:21-6.

van Wassenaer AG, Kok JH, Briet JM, van Baar A, de Vijlder JJ. Thyroid function in preterm newborns; is T4 treatment required in infants < 27 weeks' gestational age? Exp Clin Endocrinol Diabetes 1997;105(S4):12-8.

van Wassenaer, Kok JH, Dekker FW, de Vijlder JJ. Thyroid function in very preterm infants: influences of gestational age and disease. Pediatr Res 1997;42:604-9.

Smit BJ, Kok JH, de Vries LS, van Wassenaer AG, Dekker FW, Ongerboer de Visser BW. Somatosensory evoked potentials in very preterm infants in relation to L-thyroxine supplementation. Pediatrics 1998;101:865-9.

Smit BJ, Kok JH, de Vries LS, van Wassenaer AG, Dekker FW, Ongerboer de Visser BW. Motor nerve conduction velocity in very preterm infants in relation to L-thyroxine supplementation. J Pediatr 1998;132:64-9.

van Wassenaer AG, Kok JH, Dekker FW, Endert E, de Vijlder JJ. Thyroxine administration to infants of less than 30 weeks gestational age decreases plasma tri-iodothyronine concentrations. Eur J Endocrinol 1998;139:508-15.

Van Wassenaer AG, Kok JH, Briet JM, Pijning AM, de Vijlder JJ. Thyroid function in very preterm newborns: possible implications. Thyroid 1999;9(85-91).

Briet JM, van Wassenaer AG, van Baar A, Dekker FW, Kok JH. Evaluation of the effect of thyroxine supplementation on behavioural outcome in very preterm infants. Dev Med Child Neurol 1999;41:87-93.

Briet JM, van Wassenaer AG, Dekker FW, de Vijlder JJ, van Baar A, Kok JH. Neonatal thyroxine supplementation in very preterm children: developmental outcome evaluated at early school age. Pediatrics 2001;107:712-8.

Vanhole 1997

{published and unpublished data}

* Vanhole C, Aerssens P, Naulaers G, Casneuf A, Devlieger H, Van den Berghe G, de Zegher F. L-thyroxine treatment of preterm newborns: clinical and endocrine effects. Pediatr Res 1997;42:87-92.

Vanhole C, Aerssens P, Devlieger H, de Zegher F. L-Thyroxine treatment of preterm newborns. Pediatr Res 1996;40:555.

Excluded studies

Amato 1988

{published data only}

Amato M, Pasquier S, Carasso A, Von Muralt G. Postnatal thyroxine administration for idiopathic respiratory distress syndrome in preterm infants. Horm Res 1988;29:27-30.

Eggermont 1984

{published data only}

Eggermont E, Vanderschueren Lodeweyckx M, De Nayer P, Smeets E, Vanacker G, Cornette C, Jaeken J, Devlieger H, Eeckels R, Beckers C. The thyroid-system function in preterm infants of postmenstrual ages of 31 weeks or less: evidence for a "transient lazy thyroid system". Helv Paediatr Acta 1984;39:209-22.

Schonberger 1981

{published data only}

Schonberger W, Grimm W, Emmrich P, Gempp W. Reduction of mortality rate in premature infants by substitution of thyroid hormones. Eur J Pediatr 1981;135:245-53.

van Wassenaer 1993

{published data only}

van Wassenaer AG, Kok JH, Endert E, Vulsma T, de Vijlder JJ. Thyroxine administration to infants of less than 30 weeks' gestational age does not increase plasma triiodothyronine concentrations. Acta Endocrinol Copenh 1993;129:139-46.

References to studies awaiting assessment

THORN

{unpublished data only}

Markiewicz M. THORN (Thyroid hormone replacement in neonates). British Association of Perinatal Medicine Clinical Trials Group Newsletter. Summer 1998;No. 12.

References to ongoing studies

Golombek

{unpublished data only}

* indicates the primary reference for the study

Other references

Additional references

Amato 1988

Amato M, Pasquier S, Carasso A, Von Muralt GSO. Postnatal thyroxine administration for idiopathic respiratory distress syndrome in preterm infants. Horm-Res 1988;29:27-30.

Ballabio 1989

Ballabio M, Nicolini U, Jowett T, Ruiz de Elvira MC, Ekins RP, Rodeck CH. Maturation of thyroid function in normal human foetuses. Clin Endocrinol 1989;31:565-71.

Bernal 1995

Beranl J, Nunez J. Thyroid hormones and brain development. Eur J Endocrinol 1995;133:390-9.

Crowther 1998

Crowther CA, Alfirevic Z, Haslam RR. Prenatal thyrotropin-releasing hormone for preterm birth (Cochrane Review). In: The Cochrane Library. Oxford: Update Software., Issue 2, 2001. Oxford: Update Software.

den Ouden 1996

den Ouden AL. Kok JH, Verkerk PH, Brand R, Verloove-Vanhorick SP. The relation between neonatal thyroxine levels and neurodevelopmental outcome at age 5 and 9 years in a national cohort of very preterm and/or very low birth weight infants. Pediatr Res 1996;39:142-5.

Frank 1996

Frank JE, Faix JE, Hermos RJ, Mullaney DM, Rojan DA, Mitchell ML, Klein RZ. Thyroid function in very low birth weight infants: effects on hypothyroidism screening. J Pediatr 1996;128:548-54.

Franklin 1986

Franklin RC, Purdie GL, O'Grady CM. Neonatal thyroid function: prematurity, prenatal steroids, and respiratory distress syndrome. Arch Dis Child 1986;61:589-92.

Hempel 1993

Hempel MS. The neurological examination for toddler-age [PhD thesis]. The Netherlands: Universisty of Groningen, 1993.

Kooistra 1994

Kooistra L, Laane C, Vulsma T, Schellekens JMH, van der Meere JJ, Kalverboer AF. Motor and cognitive development in children with congenital hypothyroidism: a long-term evaluation of the effects of neonatal treatment. J Pediatr 1994;124:903-9.

Liggins 1988

Liggins GC, Schellenberg JC, Manzai M, Kitterman JA, Lee CC. Synergisms of cortisol and thyrotropin releasing hormone in lung maturation in fetal sheep. J Appl Physiol 1988;65:1880-4.

Linder 1997

Linder N, Davidovitch N, Reichman B, Kuint J, Lubin D, Meyerovitch J, Sela BA, Dolfin Z, Sack J. Topical iodine-containing antiseptics and subclinical hypothyroidism in preterm infants. J Pediatr 1997;131:434-9.

Lorenz 1998

Lorenz JM, Wooliever DE, Jetton JR, Paneth N. A quantitative review of mortality and developmental disability in extremely premature newborns. Arch Pediatr Adolesc Med 1998;152:425-35.

Lucas 1988

Lucas A, Rennie J, Baker BA, Morley R. Low plasma triiodothyronine and outcome in preterm infants. Arch Dis Child 1988;63:1201-6.

Lucas 1996

Lucas A, Morley R, Fewtrell MS. Low triiodothyronine concentration in preterm infants and subsequent intelligence quotient (IQ) at 8 year follow up. BMJ 1996;312:1132-3.

Meijer 1992

Meijer WJ, Verloove-Vanhorick SP, Brand R, van den Brande JL. Transient hypothyroxinaemia associated with developmental delay in very preterm infants. Arch Dis Child 1992;67:944-7.

Paneth 1998

Paneth N. Does transient thyroxinemia cause abnormal neurodevelopment in premature infants? Clin Perinatol 1998;25:627-43.

Porterfield 1993

Porterfield SP, Hendrich CE. The role of thyroid hormones in prenatal and neonatal neurological development - current perspectives. Endocr Rev 1993;14:94-106.

Radunovic 1991

Radunovic N, Dumez Y, Nastic D, Mandelbrot L, Dommergues M. Thyroid function in fetus and mother during the second half of normal pregnancy. Biol Neonate 1991;59:139-48.

Reuss 1996

Reuss ML, Paneth N, Pinto-Martin JA, Lorenz JM, Susser M. The relation of transient hypothyroxinemia in preterm infants to neurologic development at two years of age. N Engl J Med 1996;334:821-7.

Reuss 1997

Reuss ML, Leviton A, Paneth N, Susser M. Thyroxine values from newborn screening of 919 infants born before 29 weeks' gestation. Am J Public Health 1997;87:1693-7.

Rooman 1996

Rooman RP, Du Caju MVL, Op De Beeck, Docx M, Van Acker. Low thyroxinaemia occurs in the majority of very preterm newborns. Eur J Pediatr 1996;55:211-5.

Thorpe-Beetson 1991

Thorpe-Beetson JG, Nicolaides KH, Felton CV, Butler J, McGregor AM. Maturation of the secretion of thyroid hormone and thyroid stimulating hormone in the fetus. N Engl J Med 1991;324:532-6.

Touwen 1976

Touwen B. Neurological development in infancy. In: Clinics in developmental medicine. Vol. No. 58. London: William Heinemann Medical Books, 1976.

Other published versions of this review

Osborn 1999

Osborn DA. Thyroid hormone for preventing neurodevelopmental impairment in preterm infants (Cochrane Review). In: The Cochrane Library, Issue 2, 2001. Oxford: Update Software.

[top]

Data and analyses

01 Thyroid hormones versus control (All eligible studies)

Outcome or Subgroup Studies Participants Statistical Method Effect Estimate
01.01 Abnormal mental development in survivors Relative Risk [Fixed] [95% CI] Totals not selected
01.02 Cerebral palsy in survivors 1 156 Relative Risk [Fixed] [95% CI] 0.72 [0.28, 1.84]
01.03 Bayley MDI at 7-12 months in survivors 2 191 WMD [Fixed] [95% CI] -1.14 [-5.46, 3.19]
01.04 Bayley PDI at 7-12 months in survivors 2 191 WMD [Fixed] [95% CI] 0.22 [-4.80, 5.24]
01.05 Bayley MDI at 24 months in survivors 1 157 WMD [Fixed] [95% CI] -3.50 [-11.21, 4.21]
01.06 Bayley PDI at 24 months in survivors 1 157 WMD [Fixed] [95% CI] 3.10 [-3.31, 9.51]
01.07 RAKIT IQ score at 5 years in survivors 1 156 WMD [Fixed] [95% CI] -2.10 [-7.91, 3.71]
01.08 Neonatal mortality 4 310 Relative Risk [Fixed] [95% CI] 0.64 [0.361.14]
01.09 Mortality to discharge 5 354 Relative Risk [Fixed] [95% CI] 0.70 [0.42, 1.17]
01.10 Death or cerebral palsy 1 200 Relative Risk [Fixed] [95% CI] 0.70 [0.43, 1.14]
01.11 Air leak 1 44 Relative Risk [Fixed] [95% CI] 0.75 [0.19, 2.97]
01.12 Duration of mechanical ventilation (days) 3 120 WMD [Fixed] [95% CI] -2.44 [-5.41, 0.53]
01.13 CLD in survivors (oxygen at 28 days) 5 308 Relative Risk [Fixed] [95% CI] 0.96 [0.75, 1.23]
01.14 CLD in survivors (oxygen at 36 weeks) 3 227 Relative Risk [Fixed] [95% CI] 1.02 [0.64, 1.64]
01.15 Patent ductus arteriosus 5 354 Relative Risk [Fixed] [95% CI] 0.76 [0.56, 1.03]
01.16 Intraventricular hemorrhage - any grade 3 284 Relative Risk [Fixed] [95% CI] 1.20 [0.88, 1.63]
01.17 Intraventricular hemorrhage - grade 3 or 4 3 287 Relative Risk [Fixed] [95% CI] 1.07 [0.55, 2.11]
01.18 Periventricular leukomalacia 2 247 Relative Risk [Fixed] [95% CI] 0.84 [0.16, 4.38]
01.19 Necrotizing enterocolitis 2 67 Relative Risk [Fixed] [95% CI] 0.35 [0.04, 3.16]
01.20 Retinopathy of prematurity (any grade) in survivors 3 230 Relative Risk [Fixed] [95% CI] 0.93 [0.49, 1.78]
01.21 Sepsis 1 47 Relative Risk [Fixed] [95% CI] 0.51 [0.26, 0.98]
01.22 Weight gain (grams/day) 1 23 WMD [Fixed] [95% CI] 0.51 [-2.55, 3.57]
01.23 Growth head circumference (cm/week) 1 23 WMD [Fixed] [95% CI] 0.02 [-0.15, 0.19]
01.24 Growth length (cm/week) 1 23 WMD [Fixed] [95% CI] -0.06 [-0.24, 0.12]
01.25 Clinically significant behavioural problems in survivors Relative Risk [Fixed] [95% CI] Totals not selected

02 Thyroid hormones as prophylaxis (started < 24 hours age in all studies)

Outcome or Subgroup Studies Participants Statistical Method Effect Estimate
02.01 Abnormal mental development in survivors Relative Risk [Fixed] [95% CI] Totals not selected
02.02 Cerebral palsy in survivors 1 156 Relative Risk [Fixed] [95% CI] 0.72 [0.28, 1.84]
02.03 Bayley MDI at 7-12 months in survivors 2 191 WMD [Fixed] [95% CI] -1.14 [-5.46, 3.19]
02.04 Bayley PDI at 7-12 months in survivors 2 191 WMD [Fixed] [95% CI] 0.22 [-4.80, 5.24]
02.05 Bayley MDI at 24 months in survivors 1 157 WMD [Fixed] [95% CI] -3.50 [-11.21, 4.21]
02.06 Bayley PDI at 24 months in survivors 1 157 WMD [Fixed] [95% CI] 3.10 [-3.31, 9.51]
02.07 RAKIT IQ score at 5 years in survivors 1 156 WMD [Fixed] [95% CI] -2.10 [-7.91, 3.71]
02.08 Neonatal mortality 2 240 Relative Risk [Fixed] [95% CI] 0.68 [0.37, 1.25]
02.09 Mortality to discharge 3 284 Relative Risk [Fixed] [95% CI] 0.73 [0.42, 1.26]
02.10 Death or cerebral palsy 1 200 Relative Risk [Fixed] [95% CI] 0.70 [0.43, 1.14]
02.11 Air leak 1 44 Relative Risk [Fixed] [95% CI] 0.75 [0.19, 2.97]
02.12 Duration of mechanical ventilation (days) 2 78 WMD [Fixed] [95% CI] -2.28 [-5.27, 0.72]
02.13 CLD in survivors (oxygen at 28 days) 3 243 Relative Risk [Fixed] [95% CI] 0.96 [0.72, 1.28]
02.14 CLD in survivors (oxygen at 36 weeks) 2 204 Relative Risk [Fixed] [95% CI] 1.06 [0.65, 1.74]
02.15 Patent ductus arteriosus 3 284 Relative Risk [Fixed] [95% CI] 0.81 [0.56, 1.19]
02.16 Intraventricular hemorrhage - any grade 3 284 Relative Risk [Fixed] [95% CI] 1.20 [0.88, 1.63]
02.17 Intraventricular hemorrhage - grade 3 or 4 2 240 Relative Risk [Fixed] [95% CI] 1.17 [0.56, 2.42]
02.18 Periventricular leukomalacia 1 200 Relative Risk [Fixed] [95% CI] 2.00 [0.18, 21.71]
02.19 Necrotizing enterocolitis 1 44 Relative Risk [Fixed] [95% CI] 0.33 [0.01, 7.76]
02.20 Retinopathy of prematurity (any grade) in survivors 2 207 Relative Risk [Fixed] [95% CI] 0.99 [0.49, 2.01]
02.21 Clinically significant behavioural problems in survivors Relative Risk [Fixed] [95% CI] Totals not selected

03 Thyroid hormones as treatment

Outcome or Subgroup Studies Participants Statistical Method Effect Estimate
03.01 Neonatal mortality 1 23 Relative Risk [Fixed] [95% CI] 1.09 [0.08, 15.42]
03.02 Mortality to discharge 1 23 Relative Risk [Fixed] [95% CI] 0.94 [0.45, 1.92]
03.03 CLD in survivors (oxygen at 28 days) 1 23 Relative Risk [Fixed] [95% CI] 0.73 [0.15, 3.57]
03.04 CLD in survivors (oxygen at 36 weeks) 1 23 Relative Risk [Fixed] [95% CI] 0.62 [0.25, 1.56]
03.05 Patent ductus arteriosus 1 23 Relative Risk [Fixed] [95% CI] 0.36 [0.02, 8.04]
03.06 Necrotising enterocolitis 1 23 Relative Risk [Fixed] [95% CI] 0.73 [0.15, 3.57]
03.07 Retinopathy of prematurity (any grade) in survivors 1 23 Relative Risk [Fixed] [95% CI] 0.73 [0.15, 3.57]
03.08 Weight gain (grams/day) 1 23 WMD [Fixed] [95% CI] 0.51 [-2.55, 3.57]
03.09 Growth head circumference (cm/week) 1 23 WMD [Fixed] [95% CI] 0.02 [-0.15, 0.19
03.10 Growth length (cm/week) 1 23 WMD [Fixed] [95% CI] -0.06 [-0.24, 0.12]

Additional tables

  • None noted.

Amended sections

  • None noted

This review is published as a Cochrane review in The Cochrane Library, Issue 2, 1999 (see http://www.thecochranelibrary.com External Web Site Policyfor information). Cochrane reviews are regularly updated as new evidence emerges and in response to feedback, and The Cochrane Library should be consulted for the most recent recent version of the review.