Background
Congenital heart disease (CHD) is the most common congenital malformation among live births, accounting for approximately one-third of birth defects [
1]. A systematic review showed that the prevalence of CHD increased from 0.6 ‰ in the 1930s to 9.1 ‰ at the end of the 20th century [
2]. Cardiopulmonary bypass (CPB) is often used for surgery in complex CHD cases. However, hemodilution, hypothermia and ultrafiltration during CPB have been shown to induce a temporary hypothyroid state, especially in infants [
3,
4].
The thyroid hormones triiodothyronine (T3) and thyroxin (T4) have permissive effects on β1-adrenergic receptors, which enhance heart contractility and reduce systemic vascular resistance [
5,
6]. Additionally, thyroid hormones increase preload and decrease afterload, leading to increased cardiac output [
7]. Several studies have shown that the hypothyroid state, which affects the myocardial energy metabolism, is associated with poor prognosis after cardiac surgery with CPB [
8,
9]. Correspondingly, thyroid hormone replacement therapy could provide clinical benefits in infants undergoing CPB [
10‐
12].
A study was conducted to assess the effects of CPB on thyroid function in infants weighing less than 5 kg, and the results showed that low T3 and T4 levels were both predictors of high mortality [
13]. Talwar et al. [
14] found that low postoperative total T4 (TT4) levels were correlated with postoperative morbidity, a prolonged postoperative course, and prolonged mechanical ventilation (MV) in open heart surgery with CPB. Since thyroid hormone levels play a critical role in recovery from cardiac surgery and thyroid hormones decrease after CPB, the preoperative level of thyroid hormones could be a predictor of intensive care unit (ICU) mortality after CPB in CHD patients. Therefore, we conducted a retrospective study to evaluate the effect of preoperative thyroid hormone levels in relation to survival in patients after cardiac surgery with CPB.
Methods
We retrospectively reviewed the medical records of patients with CHD younger than 3 months old in our hospital between June 2017 and November 2019. This study protocol was approved by the Institutional Ethical Committees of the hospital (201912257-1). We excluded patients who were older than 90 days at the time of surgery, patients without CPB, patients with primary thyroid gland disease, and patients with trisomy 21 syndrome. Clinical data included gender; age; weight; Risk Adjustment for Congenital Heart Surgery-1 (RACHS-1) score [
15]; type of CHD; preoperative serum albumin level (normal range: 40–55 g/L); preoperative thyroid hormone levels [total T3 (TT3, normal range: 1.29–3.11 nmol/L), free T3 (FT3, normal range: 2.8–7.1 pmol/L), TT4 (normal range: 66-187.4 nmol/L), free T4 (FT4, normal range: 12.1–22 pmol/L), thyroid stimulating hormone (TSH, normal range: 0.2-5 µIU/ml)], which is a routine examination for patients with CHD in our clinic; CPB time; aortic cross-clamping (ACC) time; and ICU mortality.
Statistical analysis
The statistical analysis was performed with SPSS version 20.0 software (Chicago, IL, USA). Continuous variables are expressed as the mean ± standard deviation, while categorical variables are summarized as frequencies and percentages. Comparisons between groups were performed using an unpaired Student’s t-test for continuous variables and a χ2 or Fisher’s exact test for categorical variables. Multivariate binary logistic regression analysis was further conducted to assess the independent ICU mortality predictors. Receiver operating characteristic (ROC) curves were generated to examine the ability of variables to predict ICU mortality, and the area under the curve (AUC) was calculated from the ROC curve. Youden’s index, which maximizes the sum of the sensitivity and specificity, was used to define the optimal cut-off value. Statistical significance was defined as P < 0.05.
Discussion
To the best of our knowledge, the present study is the first clinical retrospective analysis of the predictive value of preoperative thyroid hormone levels in patients with CHD undergoing CPB. In our study of 133 consecutive patients, we found that FT3 may be an independent predictor of ICU mortality based on multivariate binary logistic regression and ROC curve analyses. Previous studies [
16] have reported that a low T3 level was an independent predictor of ICU mortality, which is consistent with our finding. However, for CHD patients, especially children, RACHS-1 scores could predict ICU mortality, length of ICU stay and duration of MV [
17‐
20]. In our study, compared with survivors, non-survivors had higher RACHS-1 scores. However, in the additional multivariate binary logistic regression analysis, the RACHS-1 score was not an independent mortality predictor. We suggest that RACHS-1 scores on the basis of CHD subtype should not consider the relationship of year, weight, and levels of thyroid hormone as confounding factors [
21]. Additionally, the small sample size of children with RACHS-1 scores of 5 or 6 may be another reason for the reduced statistical power. Although the CPB and ACC time were high in non-survivors, no significant difference was found compared with survivors, which demonstrates the improvement of cardiac surgery techniques and perfusion mode of CPB in China.
We found an interesting phenomenon in our study, namely, that all of the non-survivors were male. This could be related to the preference for sons over daughters, which is very common in China. Thus, when a child is diagnosed with complex CHD, based on the gender, operative risk and economic status of the parents, girls may not have the opportunity to undergo the operation, especially in rural areas [
22,
23]. However, gender showed significant association with ICU mortality, but considering the gender selective bias, we removed gender from the multivariate binary logistic regression.
Thyroid hormones have important effects on the cardiovascular system, such as increasing cardiac output and decreasing systemic vascular resistance, which are predictive of good outcomes [
24,
25]. However, several studies have verified that cardiac surgery with CPB induces a marked depression of thyroid hormones [
14,
26,
27]. Researchers have found that low T3 [
28] or T4 [
14] levels were correlated with postoperative morbidity in open heart surgery with CPB. A study was conducted to assess the effects of CPB on thyroid function in infants weighing less than 5 kg, and the results showed that low T3 and T4 levels were both predictors of high mortality [
13]. Since low postoperative low levels of thyroid hormones could lead to a poor prognosis, a preoperative increase in thyroid hormones may improve the prognosis. A multicenter randomized controlled trial (RCT) of T3 supplementation of patients undergoing heart surgery with CPB (TRICC) showed that T3 supplementation provides clinical advantages in patients younger than 5 months, but not in older patients [
29]. Talwar et al. [
10] performed an RCT study of perioperative oral T4 in patients younger than 6 months who underwent open heart surgery with CPB and found that postoperative thyroid hormone levels were reduced and that T4 supplementation reduced the duration of MV and ICU and hospital stays. Therefore, the preoperative level of thyroid hormones may predict the prognosis of patients with CHD undergoing CPB. Kumar et al. [
30] found that low T3 is an important marker of mortality in critically ill patients, while low T4 and TSH levels did not increase the predictability of mortality. A large-scale prospective, observational study of unselected ICU patients, found that FT3 was the most powerful and independent predictor of ICU mortality among the thyroid hormone indicators [
16]. However, Quispe et al. [
31] found that the FT3 level was not significantly different between survivors and non-survivors and was not a mortality predictor. This might have occurred because previous studies did not consider the relationship of FT3 with albumin as confounding factor, and when hypoalbuminemia was present, conversion of T4 to T3 was decreased, resulting the low FT3 levels [
32]. However, there was no significant difference in the preoperative albumin level between survivors and non-survivors in our study. Additionally, acute critically ill patients release large mounts of other hormones, such as cortisol, which has an inhibitory effect on TSH and eventually leads to low T3 levels [
33]. In our study, patients were not in stress states and had similar preoperative albumin levels, and with the exclusion of these two confounding factors, we found that the preoperative FT3 level may be a predictor of ICU mortality after CPB in infants with CHD younger than 3 months old.
Our study showed that FT3 is an independent predictor of ICU mortality because of its high AUC value. Thyroid hormones include T4, which represents the major form of circulating thyroid hormones (> 80 %), and T3, which accounts for a small portion (< 20 %) of circulating thyroid hormone and has a major biological effect on the heart. Moreover, the levels of TT3 and TT4 levels can be affected by the thyroxine-binding globulin (TBG) concentration or the binding ability of TBG, which may be affected by several drugs, including furosemide and heparin [
16]. In contrast, FT3 and FT4 were not affected by these conditions. Thus, the FT3 level may be a better predictor of ICU mortality than other thyroid hormones, which is consistent with our findings.
Some limitations exist in our study. First, this was a retrospective study with a small sample size, which limited the statistical power. Therefore, additional patients need to be enrolled, and a prospective randomized multicenter study should be conducted. Second, patients with RACHS-1 scores of 5 or 6 are rare, which reduced the ability of the RACHS-1 to predict ICU mortality in children with CHD. Finally, in a retrospective study, it is difficult to collect data on perioperative and postoperative dopamine levels and steroid use; thus, we did not consider the relationship of thyroid hormones with dopamine and steroids, which may provide more robust evidence to assess the predominant predictor.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.