Main

Thyroid hormones play a crucial role as a regulator of growth, of nervous system myelinization, of metabolism, and of organ functions(1). Critical illness, surgical procedures, and various drugs used in intensive care medicine have been reported to suppress the hypothalamic-pituitary-thyroid axis(24). Alterations of thyroid hormone plasma concentrations in critical illness have also been attributed to the euthyroid sick syndrome in adults and children(5). Whether the euthyroid sick syndrome contributes to critical illness rather than results from it has yet to be established(6). Cardiac surgery with cardiopulmonary bypass is accompanied by alterations in the CNS(7) and in many endocrine systems, but their significance remains largely unknown(8). In children with congenital cardiac malformations recovering from cardiac surgery, sequelae of the postoperative intensive care course, such as low cardiac output, left ventricular dysfunction, increased vascular resistance, and impaired ventilatory drives, resemble symptoms associated with hypothyroidism(1). Transient hypothyroidism after cardiac surgery in children may not only compromise the immediate postoperative course, but may also endanger future development of the CNS. Long-term follow-up studies on the relation between the neurodevelopmental outcome of children with congenital heart malformations and thyroid function after cardiac surgery are not available. Nevertheless, there is experimental evidence(9, 10) and clinical evidence in adult patients(11) that T3 has beneficial inotropic effects after cardiopulmonary bypass operations and may prevent low cardiac output(12).

In this study we assessed thyroid function in newborns and older children with congenital cardiac malformations before and after corrective cardiac surgery by measuring sequentially plasma TSH, T3, rT3, T4, fT4, and Tg concentrations and urinary iodine excretion. In addition, we evaluated the postoperative clinical course of all patients and correlated it with the postoperative thyroid hormone status.

METHODS

Thyroid function was studied in 132 children (80 male; 52 female) with congenital cardiac malformations (84 acyanotic; 48 cyanotic) before and every other day after (d 1 to 21) cardiac surgery for a median time of 5 d (5-17 d). The study population consisted of 22 neonates (median age 10 d; 2-28 d), 26 infants (median age 4.5 mo; 1.1-11.6), 42 toddlers (median age 3.4 y; 1-5.9 y), and 42 children (median age 12.8 y; 6.2-16.2 y). Patients with preexisting metabolic or endocrine disease were excluded from this study.

The study was approved by the Ethic Committee of the medical faculty of the University of Heidelberg.

The surgical procedure lasted 210 min (30-445 min); bypass time ranged from 18 to 204 min (median 92 min; n = 115); the core temperature was lowered to 26.6°C (17.2-34.9°C; n = 102). Circulatory arrest was required in 46 patients for a median time of 23 min (1-66 min), and aortic cross clamp was needed in 104 patients for a median time of 46 min (9-157 min).

Perioperative disinfection of the skin and mucous membranes was routinely performed with a topical anti-infective solution consisting of iodine (Braunol solution, Braun Melsungen, Germany).

After surgery 85 patients required dopamine treatment (3.9 μg/kg/min; 0.17-10 μg/kg/min) for 5 d (1-21 d). None of the patients received other drugs known to affect thyroid function.

The postoperative inotropic support with catecholamines (dopamine, dobutamine, epinephrine) and diuretics (furosemide) was calculated as the mean cumulative dose in milligrams per kg of body weight. The postoperative need for mechanical ventilation, oxygen supplementation, and duration of both intensive and regular care was assessed in days.

Blood samples were drawn between 0800 and noon from indwelling venous catheters. Plasma concentrations of TSH (Behringwerke AG Marburg, Germany); T4, fT4, and T3 (Ciba Corning Diagnostics Corporation, Medfield, USA); Tg (Henning GmbH, Berlin, Germany); and rT3 (Serono Diagnostics GmbH, Freiburg, Germany) were measured by specific commercial RIA kits, which are used for routine hormone measurements in our laboratory. A modified ceric acid reaction was used for photometric determination of urinary iodine excretion(13).

All results are expressed as median and 10th and 90th percentiles. Preoperative plasma thyroid hormone concentrations and urinary iodine excretion were compared with the maximal postoperative change and to the maximal change during recovery thereafter. In addition, the postoperative day of the maximal hormonal changes and the postoperative time interval to reach individual preoperative thyroid hormone concentrations were assessed.Statistical analysis was performed by nonparametric tests as appropriate (Wilcoxon signed rank test; Mann-Whitney U test; Friedman test; Spearman rank correlation).

RESULTS

Perioperative thyroid hormone plasma concentrations in all patients. Before cardiac surgery plasma thyroid hormone concentrations of all patients were within the normal range for age. During the postoperative period plasma levels of thyroid hormones changed considerably in all patients. TSH, T3, T4, fT4, and Tg (n = 132) plasma concentrations decreased significantly, whereas rT3 plasma levels(n = 40) increased. Maximal postoperative changes occurred irrespective of the patient's age, gender, and cardiac malformation. Therefore, these results were pooled (Fig. 1).

Figure 1
figure 1

Perioperative plasma thyroid hormone concentrations in 132 pediatric patients. rT3 plasma concentrations were measured in 40 patients. Preoperative plasma hormones levels (left plots) are compared with the maximal postoperative change (middle plots) and with the maximal change during recovery thereafter (right plots). Box plots represent the 10th, 25th, 50th, 75th, and 90th percentiles. Statistical analysis was performed by Wilcoxon signed rank test.* p < 0.0001; ** p < 0.02

Compared with preoperative plasma levels, TSH declined by 74% (22-92%), T3 by 69% (38-85%), T4 by 58% (31-87%), fT4 by 45%(19-77%), and Tg by 59% (14-93%), respectively. This represents a postoperative decrease of thyroid hormones by 58% (36-77%). In contrast, rT3 plasma concentrations rose by 69% (0-367%) after surgery. T4 plasma concentrations correlated significantly with fT4 plasma concentrations before (r = 0.75; p < 0.0001) and after(r = 0.83; p < 0.0001) cardiac surgery.

The maximal postoperative changes of TSH and rT3 were detected on d 1 (1-3 d) after surgery, whereas the postoperative trough of T3, T4, and fT4 was observed on d 2 (1-4 d) (p < 0.001), and the postoperative nadir of Tg was noticed on d 3 (1-8 d)(p < 0.0001). Five days (3-12 d) after surgery, TSH secretion increased to plasma levels above those observed before surgery: 2.2 mU/L (1-6 mU/L) versus 1.9 mU/L (0.8-3.4 mU/L) (p < 0.0001); at this time rT3 and fT4 plasma concentrations were similar to those measured preoperatively, but T3, T4, and Tg remained inferior to their respective preoperative concentrations (Fig. 1).

The effect of dopamine treatment on thyroid hormone plasma concentrations. Changes of TSH, T3, T4, and fT4 plasma concentrations after cardiac surgery were most striking in patients receiving a continuous dopamine infusion (n = 85) (Fig. 2). Plasma concentrations of rT3 were not affected by dopamine treatment. Thyroid hormone plasma levels decreased by 64% (43-80%) in patients treated with dopamine and by 47% (33-66%) in patients not treated with dopamine (p < 0.0001).

Figure 2
figure 2

Lowest plasma thyroid hormone concentrations of 132 patients after cardiac surgery. rT3 plasma concentrations were measured in 40 patients. Patients are grouped by dopamine treatment, and each group is split by cardiopulmonary bypass (CPB +, □; CPB -, □). 1) Patients treated with dopamine (n = 85) were compared with those patients not treated with dopamine (n = 47) and 2) patients, who underwent cardiopulmonary bypass operations and were treated with dopamine thereafter (n = 77), were compared with patients(n = 55), who underwent cardiopulmonary bypass not followed by dopamine treatment and those patients with or without dopamine treatment alone. Box plots represent the 10th, 25th, 50th, 75th, and 90th percentiles. Statistical analysis was performed by Mann-Whitney U test. *p < 0.0001; ** p < 0.001; *** p < 0.002

The maximal decrease of TSH, T3, T4, fT4, and Tg was detected 3 d (1-6 d) after cardiac surgery in patients treated with dopamine, and 2 d (1-2 d) after cardiac surgery in patients not receiving dopamine treatment (p < 0.0001). The rise of plasma TSH above preoperative levels on d 5 after surgery, which was observed in 88 patients (2.8 mU/L; 1.4-6.9 mU/L), was more pronounced (p < 0.05) in patients(n = 59) treated with dopamine (3.6 mU/L; 1.5-7.2 mU/L) than in those patients (n = 29) without this treatment (2.4 mU/L; 1.3-4.5 mU/L). The rise of T3, T4, fT4, and Tg after their respective postoperative nadir was significantly (p < 0.0001) delayed by dopamine treatment: d 8 (4-15 d) versus d 5 (4-9 d).

The duration of dopamine treatment after surgery (days) was positively correlated to the postoperative day, when this TSH rise occurred (r= 0.7; p < 0.0001). A weak correlation was calculated between the duration of dopamine treatment and the postoperative change of thyroid hormones (r = 0.4; p < 0.001).

The effect of cardiopulmonary bypass on thyroid hormone plasma concentrations. Those patients (n = 77), who underwent cardiopulmonary bypass operations and were treated with dopamine thereafter, were found to exhibit lower TSH, T3, T4, and fT4 plasma levels than those patients (n = 55) undergoing cardiopulmonary bypass operations not followed by dopamine treatment and those patients with or without dopamine treatment alone (Fig. 2). Postoperative plasma levels of rT3 were similar in both groups. Plasma levels of TSH, T3, T4, fT4, and Tg fell by 64% (47-82%) in patients after cardiopulmonary bypass operations followed by dopamine treatment compared with 47% (33-66%) in all other patients (p < 0.0001).

Maximal hormonal changes occurred 3 d (1-6 d) after cardiopulmonary bypass operations when patients were treated with dopamine and 2 d (1-3 d) after surgery when patients received either treatment alone (p < 0.0001).

Urinary iodine excretion (n = 105) increased from 83.4 μg/L(33.3-212.4 μg/L) before cardiac surgery to 428.9 μg/L (194.2-555.8μg/L) thereafter (p < 0.0001), representing a 370% (52-1159%) perioperative rise of urinary iodine excretion. The peak of urinary iodine excretion was measured on d 1 (1-3 d) after cardiac surgery. Subsequently urinary iodine excretion decreased, but was still elevated 5 d after surgery: 124.2 μg/L (29.9-397.5 μg/L) (p < 0.0001; compared with preoperative urinary iodine excretion). A weak negative correlation was found between maximal postoperative urinary iodine excretion and lowest postoperative plasma concentrations of T3 (r = -0.28;p < 0.01), T4 (r = 0.29; p < 0.01) and fT4 (r = -0.26; p ≤ 0.01).

Clinical course and postoperative therapy. For analysis the patients were assigned retrospectively to two groups according to the median lowest plasma T3 concentrations of the study population after cardiac surgery (0.6 nmol/L): group A (T3 < 0.6 nmol/L) consisted of 52 patients and group B (T3 ≥ 0.6 nmol/L) of 80 patients. TSH, T4, and fT4 plasma concentrations were significantly lower in group A than in group B (data not shown).

The total hospitalization period was longer in group A (23.5 d; 13.7-50.2 d) than in group B (19.5 d; 14-41.5 d; p < 0.06). This difference was probably due to a longer postoperative intensive care treatment period in group A (7 d; 1-21 d) than in group B (1 d; 1-15 d) (p < 0.0001). The treatment period on the regular cardiology ward was similar in both groups(group A, 16 d (3.5-34 d); group B, 16.5 d (12-34 d); p = NS). After surgery, group A patients required longer periods of mechanical ventilation (5 d; 1-21 d) and oxygen supplementation (10 d; 1.7-21 d) than patients of group B (1 d; 1-9.5 d; and 2 d; 1-15 d, respectively; p < 0.0001).

The postoperative medication differed remarkably between the patients of group A and group B. After cardiac surgery group A patients were treated cumulatively with 0.4 mg/kg (0-1.2 mg/kg) epinephrine, 46.7 mg/kg (0-138 mg/kg) dobutamine, 39.8 mg/kg (0-112.9 mg/kg) dopamine, and 26.6 mg/kg(0.1-62.4 mg/kg) furosemide, whereas group B patients received 0.1 mg/kg(0-0.5 mg/kg) epinephrine (p < 0.006), 14.7 mg/kg (0-54 mg/kg) epinephrine (p < 0.0001), 12.5 mg/kg (0-39.6 mg/kg) dopamine(p < 0.0001), and 15.1 mg/kg (0-44.2 mg/kg) furosemide(p < 0.003).

DISCUSSION

This study documents a remarkable suppression of thyroid function in 132 children with congenital cardiac malformations recovering from cardiac surgery and is the first study to provide evidence that thyroid dysfunction is linked to postoperative convalescence of these children, suggesting a potential impact of thyroid hormone status on the overall outcome. Moreover, our data indicate that the surgical procedure itself triggers a central inhibition of the thyroid axis, presumably at the hypothalamic-pituitary level, and that cardiopulmonary bypass operations, dopamine infusion, and probably iodine contamination augment this inhibitory effect.

Alterations of plasma thyroid hormone concentrations in nonthyroidal illness, established primarily in adult patients, have been referred to as the euthyroid sick syndrome(5). These changes have been postulated to represent an adaptive response of the organism to minimize metabolic demands during the stress of nonthyroidal illness(14). Although euthyroidism has been generally assumed in critically ill patients, convincing clinical evidence has not been provided to exclude manifest hypothyroidism. In addition, some drugs applied in intensive care medicine, such as dopamine and glucocorticoids, have been observed to produce a similar syndrome of thyroid hormone deficiency(15, 16). Allen et al.(17) suggested that changes in the level of illness severity precede, but do not follow, thyroid hormone alterations in critically ill children.

The successive changes of plasma thyroid hormone concentrations reported in our study resemble thyroid hormone alterations associated with secondary hypothyroidism. The suppression of pituitary TSH secretion after surgery was followed by a decrease of plasma T4, fT4, and T3, where plasma fT4 paralleled changes of total plasma T4. This observation is comparable to results of a recent study evaluating the effect of cardiopulmonary bypass surgery on the thyroid hormone status in 14 pediatric patients(18).

The etiology of the postoperative decrease of total T4 despite its half-life of more than a week is currently unknown. In addition to the decreased synthesis, it has been postulated that circulating inhibitors of T4 binding to plasma proteins(19) and a reduction of cellular uptake of thyroid hormones(20) cause a decline of total plasma T4 and may produce a rise of the fT4 plasma fraction, leading to increased peripheral T4 disappearance and an accelerated metabolic clearance rate(21). An altered tissue distribution of thyroid hormones and an increased demand for T4 during nonthyroidal illness may also contribute to low postoperative T4 plasma concentrations.

Furthermore, we observed a substantial decline of plasma Tg, the TSH-dependent matrix protein of thyroid hormone synthesis(22), in all patients 2 d after the decrease of TSH. These subsequent events reflect the marked difference in half-life between TSH and Tg(23), and on the other hand, these data indicate that in fact decreased thyroid hormone synthesis contributes to the low postoperative T4 and T3 plasma concentrations in our patient group. During subsequent convalescence, a TSH rebound release preceded the recovery of peripheral thyroid hormones and Tg, reflecting the recovery of the hypothalamic-pituitary-thyroid axis from transient suppression in critically ill children after surgical repair of congenital cardiac malformations. In contrast, rT3 plasma levels increased after surgery in all individuals beyond the newborn period and returned to preoperative plasma concentrations before T4 and T3 rose, indicating an altered peripheral metabolism of iodothyronines, particularly in the immediate postoperative intensive care period.

Central suppression of the thyroid axis with extremely low circulating T4, fT4, and T3 was found to be exaggerated in patients after cardiopulmonary bypass operations and in patients receiving dopamine treatment. Cardiopulmonary bypass and dopamine have been associated in adults with a blunted TSH response to thyrotropin-releasing hormone stimulation(24, 25). So far, investigations have mainly been restricted to the effect of either cardiopulmonary bypass surgery or dopamine on the thyroid hormones homeostasis in pediatric patients(15, 26, 27). In accordance with a previous study evaluating the effect of dopamine on the thyroid status of critical ill children(15), our data demonstrate that dopamine suppressed TSH secretion, but dopamine withdrawal evoked a rebound TSH release followed by an increase of T4, fT4, T3, and Tg, respectively.

Urinary iodine excretion, a measure of body iodine intake, increased after surgery in all patients studied. This finding reflects the perioperative application of a topical anti-infective agent containing iodine. Transcutaneous iodine absorption has been well established in children undergoing cardiac surgery(28). Especially young children are extremely sensitive to iodine-induced hypothyroidism, because the mechanisms for decreasing thyroid iodine uptake to compensate for high plasma iodine levels are immature(29). Our study failed to directly prove the suppressive effect of iodine because no adequate control group was available. However, iodine administration to normal individuals has been shown to block the response of the thyroid gland to TSH stimulation, resulting in reduced T4 plasma levels with elevated TSH plasma levels(30). Because our data demonstrated decreased rather than increased plasma TSH concentrations after surgery, it is likely that the hypothalamic-pituitary-thyroid axis was suppressed postoperatively.

Our data suggest a multifactorial etiology consisting of endogenous and exogenous factors that lead to the reported alterations in the thyroid hormone status of pediatric patients recovering from cardiac surgery. Low postoperative plasma iodothyronine concentrations occurred in all children studied, irrespective of the surgical procedure and postoperative medication. These changes may have been triggered by the endogenous release of mediators such as glucocorticoids(31, 32), tumor necrosis factor(33), and cytokines such as IL-6(34), that are known to peak after cardiopulmonary bypass operations in adults and to suppress thyroid function(35). Exogenous factors, such as dopamine infusion and iodine skin preparations, further aggravate the hypothyroid state after surgery. The postoperative clinical course of our patients was clearly linked to their postoperative thyroid hormone status. Low circulating T4 and T3, observed 1-2 d after surgery, correlated well with the postoperative need for vasoactive and inotropic catecholamines, diuretics, and duration of mechanical ventilation and oxygen supplementation. The intensive care treatment period was significantly prolonged in patients with low postoperative T3 plasma concentrations. The suppression of the thyroid axis was not correlated with the postoperative mortality reported in adult patients(36, 37). In our study plasma T4 and T3 concentrations as low as 5.2 and 0.3 nmol/L, respectively, were measured after cardiac surgery.

This study provides strong evidence for the concept that alterations in thyroid function in critically ill children after cardiac surgery are evoked by suppression of the hypothalamic-pituitary axis, by decreased thyroid gland hormone production, and to a minor extent by altered thyroid hormone metabolism. Further studies concerning the endocrine responses to cardiac surgery in pediatric patients are warranted to improve our understanding of the specific postoperative pathophysiology, before hormone replacement therapy can be advocated.