Risk factors of hypothyroxinemia in premature infants

Background: Hypothyroxinemia is dened by low levels of thyroxine (T4) despite low or normal levels of thyroid-stimulating hormone (TSH). This study aimed to evaluate the risk factors associated with transient hypothyroxinemia (THOP) in premature newborn. Method: This is a single center, retrospective, case-control study. Premature newborns, between 24-34 weeks of gestation, hospitalised between January 2014-December 2019 in our Newborn Intensive Care Unit (NICU) were analyzed through their medical records. Thyroid function tests were routinely performed between the 10th and 20th days of postnatal life and were evaluated according to the gestational age references. Thirty six risk factors (prenatal and postnatal parameters, medical treatments, clinical diagnoses and applications in NICU) were searched in the patient group with THOP (n=71) and the control group with euthyroid prematures (n=73). The risk factors for THOP were identied by univariate analysis, followed by multivariate analysis. Results: Mean gestational ages of the study and the control groups were 29.7 ± 2.48 and 30.5 ± 2.30 weeks, respectively (p = 0.606). The birth weight, small for gestational age (SGA), intraventricular hemorrhage (IVH), congenital heart disease (CHD) were found to be the risk factors for THOP in the univariate analysis and CHD (p=0.033, odds ratio [OR]:3.7, 95% condence interval [CI]: 1.1-12.3), BW (p=0.012, OR:0.998, 95% CI: 0.9-1.01) and SGA (p=0.006, OR:5.3, 95% CI: 1.6-17.71) were found to be an independent risk factor for THOP as a result of the multivariate analysis. According to multivariate analysis, birth weight, SGA and detection of CHD were found to be statistically significant (p<0.05, dark stained ones).


Background
Transient hypothyroxinemia of prematurity (THOP) is de ned by low levels of thyroxine (T4) despite low or normal levels of thyroid-stimulating hormone (TSH) (1). Hypothyroxinemia is observed in around 50 % of premature newborns and its risk increases as the gestational week decreases (2,3). Serum T4 and free T4 (FT4) levels in premature newborns vary according to gestational age in the rst days of life. T4 and FT4 concentrations decline to the lowest level between ten and fourteenth days after birth. This situation is more severe with low gestational week and birth weight (BW) (4). In term infants (37-42 weeks' gestation) serum T4 levels characteristically increase in the rst week of life whereas in infants born prematurely, and especially those below 30 weeks' gestation, may decrease transiently resulting in a period of hypothyroxinaemia (5,6).
Postpartum TSH and thyroid hormone peaks observed in term newborns are not so evident in preterms.
Especially below 30 weeks' gestation, TSH increase is late and weak, while T4 and FT4 levels remain low (7). In premature newborns, physiologicaly low thyroid hormones can be explained by: blunted phsyiolgic hyperthyroidim (correlated with gestational age), increase of thyroid hormone demand (thermogenesis, cardiac and skeletal muscle functions), de ciencies in iodine metabolism, immaturity of hypothalamopituitary-thyroid axis, insu cient response of thyroid gland to TRH, low T4 conversion from T3 and early interruption of maternal T4 transport (8,9). In addition to these factors, thyroid functions may also be suppressed due to medications that are commonly used in premature infants (dopamine, dexamethasone), disorders, such as respiratory distress syndrome, infections, necrotizing enterocolitis, patent ductus arteriosus, malnutrition, chorioamnionitis, as well as iodine de ciency or excess (8, [10][11][12]. Thus THOP can be observed due to multiple factors in premature newborns. This study is one of the most important series in the literature which aimed to evaluate the risk factors associated with THOP (4,11,12).

Participants and datas
This is a single center, retrospective, case-control study. THOP and control groups (euthyroid) of premature newborns who had been matched according to their gestational age were compared. File records of 538 newborns who had been admitted to Istanbul University-Cerrahpaşa Faculty of Medicine Neonatal Intensive Care Unit (NICU) between January 2014-December 2019 were retrospectively analyzed. Recruitment criteria: premature newborns between 24-34 GA, without multiple major congenital anomaly and whose serum thyroid function test were done between the 10th and 20th days of postnatal life. In the 5-year retrospective examination: THOP, euthyroid, primary hypothyroidism and subclinical hypothyroidism were diagnosed in 45.8% (n=83), 47.5% (n=86), 4.9% (n=9) and 1.6% (n=3), respectively.
Infants of mothers with maternal hypothyrodism, pregnancy without follow-up was excluded from both THOP and euthyroid groups. Two groups were formed as the patient group (n=71, THOP) and the control group (n=73, euthyroid). The study design was planned as shown in Figure 1.
Thirty six risk factors detected between birth to the time of blood samples collection (prenatal and postnatal parameters, medical treatments, clinical diagnoses and applications in NICU) were compared between the patient group with THOP (n=71) and the control group with euthyroid infants (n=73).
Ethics approval was obtained from Istanbul University-Cerrahpaşa Faculty of Medicine Ethics Committee (reference no: 36423).
THOP was de ned as a low FT4 (as per the age reference interval) and low or normal TSH (as per the age reference interval) levels (8). Serum TSH and FT4 measurements were performed between the 10th and 20th days of postnatal life.
Bronchopulmonary dysplasia (BPD) was de ned as the ongoing need for oxygen after 28 days (13).
Echocardiography evaluation was performed by pediatric cardiologist, patent ductus arteriosus (PDA), and other congenital heart diseases (CHD) were recorded. Premature newborns who received a mechanical ventilation or non-invasive ventilation support longer than 24 hours were considered to have these risk factors. Sepsis was diagnosed with positive blood cultures. Severe intraventricular hemorrhage (IVH) was de ned as grade 3 or 4, while the presence of severe retinopathy of prematurity (ROP) was de ned as grade 3 and beyond based on the International ROP classi cation (14,15). Fetal growth restriction (FGR) is de ned as the failure of the fetus to achieve its genetically determined growth potential (16). The cases whose weight was below the 10th percentile in antenatal percentile follow-ups were de ned as FGR by the perinatology department. Small for gestational age (SGA) is de ned by birth weight below the 10th percentile for gestational age (17). Central catheterisation was performed in two ways: umbilical arterial/venous cathether and Peripherally Inserted Central Catheter (PICC).

Statistical methods
The SPSS v.21 (SPSS Inc., Chicago, IL, USA) software was used for the statistical analyzes. The compatibility of the data with normal distribution was evaluated by the descriptive statistics (mean, standard deviation) and Kolmogorov Smirnov Test. Continuous variables were shown with a median (25th -75th quarters), and categorical variables with frequency and percentage (%). The Mann-Whitney U test was used to compare the groups in terms of continuous variables. The Chi-squared test and Fisher's exact test were used where appropriate in the analysis of the categorical variables. The variables that constituted signi cant results in the analyses were assessed through a correction made in the singlevariable logistic regression analysis based on age, gender and weight. The results were provided with the Odds Ratio (OR) and 95% con dence intervals. The level of signi cance was taken as p<0.05.
In the THOP group, a total of 15 cases of CHD; 8 cases atrial septal defect (ASD), 3 cases ventricular septal defect (VSD), 1 case transposition of the great arteries (TGA), 1 case coarctation of the aorta (CoA), 1 case atrial septal aneurysm, 1 case mitral valve prolapse were detected. In the control group, 4 patients had ASD and 1 VSD. In THOP group, CHD was statistically signi cantly higher than control group (p=0.018, Table 2).

Discussion
In this study, THOP and control groups of newborns who had been matched according to their gestational age were compared. Low birth weight, SGA and CHD were found to be as independent risk factors. These three risk factors are not affected by the gestational age, the severity of the illness and the medical and clinical treatment practices applied. It is already shown in the literature that the frequency of hypothyroxinemia changes in proportion to gestational age (18).

THOP and SGA
In our study, interestingly although THOP group had similiar gestational age with the control group, had a low birth weight in THOP group. Consequently in this study, incidence of SGA was higher in the THOP group and increased the risk of the THOP 5.3 times, one hundred gram increase in BW reduced the THOP risk by 20%. We speculate that being SGA may result in further de cit in storage and adaptive mechanisms. A recent study by Chunhua et al, supported our results that SGA in preterms may be associated with thyroid dysfunction (19). A comprehensive study by Bagnoli et al showed that similar to our results, preterm SGA neonates had lower FT4 compared to preterm AGAs and TSH levels were similar (20). It is generaly accepted that hypothyroxinemia is usually transient in preterm SGAs and is caused by placental hypoxia and delayed maturation of the thyroid gland (21). In addition, some studies underlined that hypothyroxinemia in preterm SGAs might be attributed to nutritional de ciency and might be reversible with the regulation of nutrition (22,23). Similar to the literature, the results of our study suggested that thyroid function tests should be followed more closely in both preterm and SGA newborns.

THOP and CHD
It is well known that thyroid and cardiac disorders can be associated. Congenital heart diseases (5.5%) are frequently associated with congenital hypothyroidism sugggesting common genetic mechanisms involved in thyroid and heart development (24). The association of CHD and thyroid disorder in Down syndrome, which is a genetic disorder, has been well de ned (25). In the recent study of HJ Lee et al., it was found that the coexistence of CHD and transient thyroid disorders is approximately 50% (26). In a study with 76 preterm infants hospitalized in neonatal units, cardiovascular disease was signi cantly higher in THOP (27). In a comprehensive study by Sadia Malik et al., evaluating the relationship between SGA and CHD, it was found that the risk of being SGA was twice as high as in the control group (28). Decreased birth weight, SGA and CHD may have common pathogenetic mechansims that are associated with THOP. Similarly, in this study, hypothyroxinemia was found with a higher rate in preterms with SGA and CHD. It is well known in the literature that the combination of CHD and SGA increases mortality and morbidity (29). Therefore, close follow-up of thyroid function tests is more important in preterms with CHD and SGA.

THOP and prenatal-postnal conditions
Some studies have shown that in preterms thyroid functions are affected by postpartum drugs and some perinatal conditions. Drugs frequently used in premature infants (dopamine, dexamethasone), respiratory distress syndrome, infections, disorders such as necrotizing enterocolitis, patent ductus arteriosus, malnutrition, chorioamnionitis, iodine de ciency or overload may suppress thyroid functions (11,(30)(31)(32)(33). It should be noted that gestational age is inversely correlated with the severity of problems in the prematüre newborns which may be associated with more intense medical and invasive treatments. Thus, it is di cult to evaluate whether a risk factor is the consequence of immaturity or the medical treament. Previous studies reported that surfactant, dopamine, glucocorticoids and erythrocyte transfusion increase the risk of THOP (34)(35)(36)(37). Most of these studies evaluating the risk factors for hypothyroxinemia belong to the past years and are studies with a smaller number of cases compared to our study.
The medications are used widely as gestational age decrease and severity of the disorder increase. In a study of very low birth weight newborns, it was shown that the negative effect of dopamine and dobutamine use on thyroid function rapidly resolved after treatment was discontinued (38). There are some studies showing that the suppression of thyroid functions has decreased in parallel with the decrease in RDS severity due to the development of prenatal care and neonatal units, especially the widespread use of antenatal steroids and early surfactant treatment in recent years (11). In a study by Lay et al. In recent years, no relationship was shown between THOP and PDA, IVH, antenatal steroid use, ROP, Apgar scores, and sepsis (39). In different pathologies like respiratory distress syndrome, PDA, sepsis, intracranial hemorrhage and necrotizing enterocolitis, it is claimed that their effects on serum thyroid hormone levels are mediated in part by acute in ammatory cytokines (11). In our study, differently from previous studies RDS, PDA, sepsis, IVH, dopamine, dobutamine and erythrocyte suspension transfusion were not found to be the risk factors for THOP. The reason for these different results in our study, may be due to less needed treatments and their temporary effects with the widespread use of surfactants and antenatal steroids in recent years and the decrease in the severity of diseases in neonatal units.
In this study, serum TFT results between ten and twenty days of life were evaluated according to the postnatal and gestational age reference values, which were generally accepted in the literature and in the largest study on this subject (1). However, there is no TFT level de ned as "low or normal" and there is no consensus on the timing of measurements. The strength of this study is that it is one of the few studies (7, 10, 34-38) evaluating so many parameters affecting thyroid functions in premature newborns. On the other hand our study have some limitations as being it is a single center and retrospective study. A multi center and prospective study could create more meaningful results.

Consent for publication
Not applicable.

Availability of data and materials
All data generated or analysed during this study are included in this published article and its supplementary information les.

Funding
There is no funding.

Authors' contributions
The authors declare that they have no competing interests. Data are given as numbers and (%), after decimal rounded to the greater side. Other data are given as mean ± standard deviation (min-max). When the THOP group and control group cases were compared, FT4 was lower as expected (p<0.05, dark stained).
Abbreviations: THOP, transient hypothyroxinemia of prematüre; GA, gestational age.  According to multivariate analysis, birth weight, SGA and detection of CHD were found to be statistically significant (p<0.05, dark stained ones).