In the 30–60 min post birth, due to the colder environment and umbilical cord clamping, the production of TSH by the infant increases to 60–80 mIU/L to then decrease to 20 mIU/L at 24 h and post birth and 6–8 mIU/L by approximately 1 week post birth.7 There is also an increase in T4 and FT4 levels (to 10–22 µg/dL and 2–5 ng/dL, respectively) at 24–36 h post birth. Levels of T3 also increase due to the increased secretion and conversion of T4 to T3 in tissues.

The levels of T4, FT4 and T3 decrease gradually in the first 4 weeks post births to 7–16 µg/dL of T4, 0.8–2 ng/dL of FT4 and 0.5–6 mIU/L of TSH.

Following birth, PT infants also experience a similar increase in TSH and thyroid hormone levels, but of lesser magnitude compared to term infants. The levels are proportional to the weeks of gestation and birth weight5,8,9 (Fig. 1).

Murphy et al.8 analysed the changes in the hypothalamic-pituitary-thyroid (HPT) axis in the first 24 h post birth in PT infants born at 24 and 34 weeks of GA and compared them based on GA. They found that the TSH peak that occurs around birth was attenuated in the group born at 24–27 weeks; the levels of T4 also declined in the first 24 h in this group, while it increased in the more mature group. As occurs in term infants, T4 values decrease in the first week of life, but this decrease is greater in PT and VLBW infants, as clearance of T4 is quicker. There are several possible causes of impaired thyroid function in PT infants, which we present in Table 2.

Normal values in PT infants relative to GA and postnatal age have yet to be established,10 which may result in overdiagnosis or underdiagnosis of hypothyroidism in these patients.

Carrascosa et al.11 analysed thyroid function in 75 healthy PT infants born between 30 and 35 weeks of gestation through the first year of life in comparison to term infants of the same postnatal age. The mean TSH concentration at 24 h post birth in PT infants was significantly lower. Also, the levels of T4 and T3 at 1 and 24 h were significantly lower and the level of reverse triiodothyronine (rT3) at 24 h significantly higher in PT infants. From 1 week post birth, the values obtained in thyroid function tests were in the same range in both groups.

Thyroid function may be abnormal in PT infants with hyaline membrane disease or respiratory distress, manifesting as euthyroid sick syndrome.12 The expected increase in TSH, T4 and T3 concentrations at birth does not occur, and levels may not increase until the child recovers and will do at a very slow pace. There is also an inverse correlation between FT4 levels and disease severity in these infants.

Infants born small for gestational age also exhibit particular alterations of thyroid function. Their TSH levels increase at birth, but stay within the normal range, and they have greater thyroid hormone requirements in the long term, similar to preterm infants, and therefore should be monitored at regular intervals.13

In short, the maturation of the HPT axis in PT and SGA newborns with an increase in TSH levels occurs between 2 and 6 weeks post birth. Despite these differences and the presence of factors that affect thyroid function in premature infants, the cut-off values used for diagnosis of CH are the same that are applied in term infants, which increases the probability of false negatives in screening, as noted above.

Preterm infants may experience delayed TSH elevation, in which case newborn screening would not detect the problem. There is evidence that the results of newborn screening may be normal in 5%–10% of VLBW or ELBW infants.14

Delayed TSH elevation in PT neonates is not well understood, and whether it is a transient disorder due to the immaturity of the axis or constitutes a moderate but persistent form of CH continues to be debated.

There are international recommendations for the detection of CH in these cases, among which we ought to highlight the following:

• –

  In the document published jointly by the American Academy of Pediatrics, American Thyroid Association and Lawson Wilkins Pediatric Endocrine Society in 2006,15 the authors acknowledged that delayed TSH elevation is more frequent in preterm infants but also discussed the difficulties of implementing a universal screening programme with routine testing of a second specimen and expressed the need of longitudinal studies to assess the long-term outcomes of these measures. An alternative option would be to limit this screening approach to patients at high risk of CH, such as VLBW infants (incidence of CH, 1 case per 250 births), ELBW infants (incidence of CH, 1 case per 1589 births) or newborn infants admitted to the neonatal intensive care unit or with cardiovascular disease, with re-evaluations at 2 and 6 weeks. In case of persistent hyperthyrotropinaemia at 6 weeks, the guideline suggests initiation of treatment and retesting at 3 years.

• –

  The consensus guidelines of the European Society of Paediatric Endocrinology published in 201416 recommended a second-screening approach in the following cases: preterm neonates with a GA of less than 37 weeks; LBW and VLBW neonates; ill and preterm neonates admitted to the neonatal intensive care unit; specimen collection within the first 24 h of life; and multiple births, particularly in cases of same-sex twins. They recommended collection of the second specimen at 2 weeks post birth or 2 weeks after the initial screening. Still, the authors recognised that this approach is not implemented in every neonatal unit and underscored the difficulty of implementing the recommendation of retesting with venous blood specimens instead of dried blood spot specimens.

Data have been obtained through the implementation of specific programmes for rescreening of thyroid function in PT infants.17 The observed frequency of CH was higher in PT infants compared to term infants, with an incidence of 1 per 579 PT births at 32–36 weeks of gestation versus 1 per 1488 term births at 37 or more weeks of gestation. The incidence was higher in infants born at or before 32 weeks: 1.56% in low birth weight infants, 1.9% in VLBW infants and 3.7% in ELBW infants.

McGrath et al.18 collected whole blood samples weekly through 37 weeks of corrected age or hospital discharge. The authors highlighted that 27 (50.9%) PT infants born before 33 weeks of gestation that received a diagnosis of CH had delayed TSH elevation detected between 8 and 48 days post birth (mean, 13 days) that would not have been detectable in the initial screen. Of these patients, 12 (40.7%) had decompensated hypothyroidism (FT4 < 10 pmol/L) and 4 had severe hypothyroidism (FT4 < 5.5 pmol/L). This study also analysed repeat screening, which identified 6 cases (22%) of permanent CH and 8 (29%) of transient CH. In 13 neonates (48%) elevation of TSH was detected after 15 days, and 7 of them had a FT4 concentration below 10 pmol/L. In addition, 25% of infants with delayed TSH elevation had a history of iodine exposure, so the authors recommended monitoring for up to 1 month after exposure.

Kaluarachchi et al.19 conducted screening in heel prick blood samples at 2 weeks, 4 weeks and discharge, with testing of venous samples performed for confirmation. They diagnosed CH in 49 PT infants, of who 92% had delayed TSH elevation. Testing of samples obtained at 2 weeks identified the most cases (n = 18), and it is worth highlighting that 15 patients had TSH concentrations greater than 100 mU/L and that 1 patient with early TSH elevation and 19 with delayed TSH elevation had free T4 concentrations of less than 0.8 ng/dL, which seems to justify monitoring with retesting at different time points.

In short, given the high incidence of abnormalities in these patients and the potential impact on their outcomes, it seems reasonable to recommend reassessment of thyroid function and monitor the patient, as the long-term impact of diagnosis or treatment in these cases is unknown.

Transient hyperthyrotropinaemia can be due to different factors: maternal thyroid disease (maternal treatment with antithyroid agents, transfer of maternal TSH receptor antibodies [TRAB]), gene variants (heterozygous variants in DUOX-2 gene or the TSHR gene that encodes the TSH receptor), prenatal/postnatal exposure to high doses of iodine (povidone-iodine, iodinated contrast media), areas of natural iodine deficiency, factors related to disease severity or use of drugs listed in Table 3.

In recent years, improvements in treatment and the management of perinatal disease (antenatal steroid therapy, noninvasive ventilation, decreased used of pharmaceuticals etc.) has achieved a decrease in the incidence of hypothyroxinaemia of prematurity and delayed hyperthyrotropinaemia in extremely PT infants.20

Persistent elevation past 2 weeks post birth with a TSH concentration greater than 10 mU/L or a FT4 concentration of less than 0.8 ng/dL is an indication for treatment recognised in most guidelines. There is less consensus as regards the approach to intermediate TSH values between 6 and 10 mU/L, which will depend on several factors, and in these cases decision-making is shared by the clinician and the parents. Based on the consensus guidelines of the European Thyroid Association, infants with persistent TSH concentrations greater than 10 mU/L after 1 month post birth are eligible for treatment through age 3 years with subsequent reassessment.20 Thyroid imaging (ultrasound or 131I o 99Tc scintigraphy) is recommended to assess for the presence of structural anomalies that would support a diagnosis of permanent CH. The identification of genetic changes associated with hyperthyrotropinaemia can also guide the decision to treat and predict the course of disease.

Preterm neonates are more likely to develop hypothyroxinaemia (low T4/FT4, normal TSH), which is detected in the first weeks of life in up to 50% of those born before 28 weeks of GA and more marked the more immature or severely ill the infant is (Table 4).21

Most cases are transient, but in the early stages it may be challenging for the clinician to discern whether the hypothyroxinaemia represents a secondary or tertiary form of hypothyroidism (hypothalamic-pituitary) or is due to TBG deficiency. The current scientific evidence is also insufficient to prove that levothyroxine for treatment of hypothyroxinaemia in PT infants can improve long-term neurodevelopmental outcomes. There are few randomised trials, conducted in small samples (10–100 preterm infants), with heterogeneity in the characteristics of participants and the protocols used (doses of levothyroxine ranging from 4 to 20 µg/kg/day, duration ranging from 2 to 6 weeks), which hinders interpretation of the data.22 A recent review23 analysed 20 articles on the impact of hypothyroxinaemia on neurodevelopment and 7 randomised trials on hormone replacement therapy with levothyroxine conducted between 1981 and 2016. Some of the evidence suggests that hormone replacement could be beneficial for extremely PT infants. Other studies have found an association between T4/T3 levels and patient outcomes (mortality, cardiovascular complications, etc.), although it would be difficult to establish causality.24,25

Treatment with thyroxine remains controversial,26–30 and the current evidence does not support recommending its routine use of treatment of PT infants with transient hypothyroxinaemia. More data is required to identify cases in which treatment may be beneficial,30 the best time to initiate treatment and its ideal duration based on the optimal cut-off points for circulating thyroid hormone levels.

We would recommend initiation of treatment in case of hypothyroxinaemia associated with an elevated TSH concentration greater than 10 mU/L or persistent elevation (FT4 < 0.8 ng/dL in 2 measurements taken 1–2 weeks apart) with the decision to treat made on a case-by-case basis in patients at high risk, such as PT infants born before 28 weeks’ gestation or with weights of less than 1000 g, especially those with severe illness (Table 5).

It is important to differentiate transient hypothyroxinaemia of prematurity from hypothyroxinaemia associated with abnormalities in the hypothalamic-pituitary axis. Persistent low FT4 levels combined with levels of TSH below or in the lower range of normal is indicative of central CH. It is most frequently associated with other pituitary hormone deficiencies (congenital panhypopituitarism) and may manifest with prolonged neonatal hypoglycaemia, micropenis or bilateral cryptorchidism and prolonged jaundice, in addition to abnormal pituitary gland morphology (ectopic neurohypophysis, anterior pituitary agenesis or hypoplasia, pituitary stalk anomalies, or midline anomalies such as septo-optic dysplasia). There is also the possibility of isolated central hypothyroidism, which is infrequent, with an incidence of 1 per 30 000 births, and cannot be detected by newborn screening of TSH levels.