Werner & Ingbar's The Thyroid: A Fundamental & Clinical Text, 9th Edition

11D.Nonthyroidal Illness

William M. Wiersinga

Changes in the function of the hypothalamic–pituitary–thyroid axis and in thyroid hormone transport and metabolism are common in patients with nonthyroidal illness. Illness in this context comprises virtually all nonthyroidal disorders, surgical and nonsurgical trauma, and inadequate caloric intake. Many patients with nonthyroidal illness also receive drugs that affect thyroid hormone regulation and metabolism, but for the sake of clarity pharmacologic interference is not considered an intrinsic part of the spectrum of changes in hypothalamic–pituitary–thyroid function that occurs in these patients.

The effects of illness are not confined to the thyroid axis; in fact, illness induces temporary changes in many neuroendocrine systems (1). The changes in thyroid function are therefore only one component of the neuroendocrine response to illness and can be viewed as part of the general adaptation to stress. It is thus reasonable to assume that these changes contribute to the maintenance of homeostasis in patients with any illness.

The changes in hypothalamic–pituitary–thyroid function are sometimes referred to as the sick euthyroid syndrome. This name is not appropriate, for several reasons. First, there is no single syndrome, but rather a constellation of changes of widely varying magnitude. Second, it is not clear that the patients are in fact euthyroid in those tissues that are targets for the action of thyroid hormone (2,3). The changes to be described here in humans and animals have been reviewed extensively (3,4,5,6,7); those that occur in animals are discussed here only to clarify aspects of the changes for which data from humans are lacking. Although these changes are heterogeneous, as is most evident from the particular changes in patients with specific disease entities, the majority of illnesses induce a rather uniform pattern of changes.

DESCRIPTION OF NONTHYROIDAL ILLNESS

Changes in Serum Thyroid Hormone and Thyrotropin Concentrations

Serum Triiodothyronine and Reverse Triiodothyronine Concentrations

A fall in serum triiodothyronine (T3) concentrations and a rise in serum reverse triiodothyronine (rT3) concentrations are the most common changes in patients with nonthyroidal illness, which is often referred to as the low T3 syndrome. Fasting induces these changes in serum T3 and rT3 concentrations within 24 to 36 hours, and the values return rapidly to baseline upon refeeding. The composition of the food is relevant: refeeding with glucose is more effective in restoring normal values than refeeding with equicaloric amounts of protein or fat (4). Similarly, serum T3 concentrations fall and serum rT3 concentrations rise within a few hours after initiation of general anesthesia and surgery, and return to normal in several days if the postoperative course is uncomplicated.

The low serum T3/high rT3 pattern is found in patients with most acute and chronic illnesses, including infectious diseases, infiltrative and metabolic disorders, cardiovascular diseases, pulmonary diseases, gastrointestinal diseases, cancer, burns, and trauma. Serum rT3 concentrations increase to supranormal values in patients with mild illnesses and do not increase much more in patients with more severe illnesses. In contrast, serum T3 concentrations decrease further as the severity of illness increases (Fig. 11D.1). Serum T3 concentrations are low or even undetectable in patients with critical illnesses, and they are persistently low in patients with chronic illnesses. The severity of the illness is in general reflected by the magnitude of the changes in serum T3 and rT3 concentrations. For example, there is an inverse relationship between serum T3 concentrations and glycosylated hemoglobin values in patients with diabetes mellitus and between serum T3 concentrations and serum creatinine concentrations in patients with renal insufficiency. Similarly, in patients who have a myocardial infarction, the size of the infarction is inversely related to the fall in serum T3 concentrations and directly related to the rise in serum rT3 concentrations.

FIGURE 11D.1. Serum thyroid hormone concentrations (filled circle, triiodothyronine (T3); open circle, reverse T3 (rT3); open triangle, thyroxine (T4); upside down open triangle, free T4 index; closed triangle, free T4, measured by equilibrium dialysis) in 504 newly hospitalized patients, divided as follows: group I, all values within the normal reference range; group II, serum rT3 values high but serum T3 and T4 values normal; group III, serum T3 values low but serum T4 values normal; group IV, serum T3 and T4 values low. The hospitalization days are given as median days for the survivors. (From Docter R, Krenning EP, de Jong M, et al. The sick euthyroid syndrome: changes in thyroid hormone serum parameters and hormone metabolism. Clin Endocrinol (Oxf) 1993;39:499, with permission.)

In general, serum free T3 concentrations fall to a lesser extent than do serum total T3 concentrations, but the results of measurements of serum free T3 in patients with nonthyroidal illness depend greatly on the assay used. Indirect methods in which the fraction of serum T3 that is unbound is measured give low values (8,9) more often than when serum free T3 is measured directly (direct measurement of free T3 after physical separation from bound T3) (10,11).

Serum Thyroxine and Free Thyroxine Concentrations

Serum thyroxine (T4) concentrations do not change during fasting in healthy subjects, but they decrease in patients with protein-calorie malnutrition. Patients with mild to moderate nonthyroidal illnesses usually have normal serum T4 concentrations. Some severely ill patients have low serum T4 concentrations (Fig. 11D.1), and among them, serum T4 concentrations are inversely correlated with the mortality rate.

There is uncertainty about serum free T4 concentrations in patients with nonthyroidal illness; the results vary depending on the method of measurement (see Chapter 13). When measured directly, using methods that involve initial physical separation of free T4 from protein-bound T4 by equilibrium dialysis or ultrafiltration with minimal dilution of serum so as not to alter the equilibrium between free and bound T4, and quantification by radioimmunoassay, serum free T4 concentrations are usually normal or even high in patients with nonthyroidal illness, even in those who are critically ill and have low serum total T4concentrations (11). In contrast, the serum free T4 values are normal or low when measured indirectly, by two-step immunoassays in which free T4 is immunoextracted from serum, analogue immunoassays that use labeled analogues of T4 that bind to T4 antibodies but not to serum thyroid hormone–binding proteins, or index methods in which free T4 values are calculated as the product of the serum total T4 concentration and a test of the number of unoccupied serum protein-binding sites (thyroid hormone–binding ratio or T3-resin uptake) (Fig. 11D.1) (12,13,14). None of these methods for measuring serum free T4 is completely free of methodologic artifacts.

The issue is further complicated by the decrease in serum protein concentrations (especially the concentrations of transthyretin and albumin) that occur in some patients with severe nonthyroidal illnesses. The analogue methods may give falsely low values for serum free T4 due to differences in serum proteins between patients with nonthyroidal illness and the standard serum used in the assays. The two-step serum free T4 assays are better validated, but vulnerable to an in vitro effect of heparin (which falsely raises measured serum free T4 values) or the influence of sample dilution (which attenuates the effect of inhibitors of binding). Furthermore, the serum of some patients with nonthyroidal illness contains substances that inhibit the binding of T4 to serum binding proteins, which may explain in part results suggesting that T4 binding is decreased in some of the patients (15,16). Based on the most reliable assays, it appears that most patients with nonthyroidal illness, even critically ill patients, have serum free T4 concentrations that are within the normal reference range.

Serum Thyrotropin Concentrations

Serum thyrotropin (TSH) concentrations are normal in most patients with nonthyroidal illness, but may increase to slightly above the normal range (10 to 20 mU/L) for a few days soon after recovery begins (17,18). However, some critically ill patients who have low serum T4 concentrations also have low serum TSH concentrations (19), although the serum TSH concentrations are rarely as low as in patients with thyrotoxicosis. The serum TSH response to thyrotropin-releasing hormone (TRH) is usually proportional to the baseline serum TSH concentration. Thus, critically ill patients with low basal serum TSH concentrations have low serum TSH responses to TRH, and both findings are indicative of a poor prognosis (20). Additionally, critically ill patients are often treated with infusions of dopamine or high doses of glucocorticoids, both of which lower serum TSH concentrations, even in normal subjects (see section on regulation of thyrotropin secretion in Chapter 10).

Changes in Thyroid Hormone Kinetics

Kinetic studies of thyroid hormone production and metabolism have been conducted in normal subjects during fasting and in patients with chronic renal failure, cirrhosis, diabetes mellitus, mild illness, and critical illness. The results can be summarized as follows (21). In patients with low serum T3 concentrations, the production rate of T3 is decreased, but its metabolic clearance rate is unchanged. The decrease in T3 production is due to decreased extrathyroidal deiodination of T4 to T3, which normally provides ~80% of the daily T3 production (Fig. 11D.2). The fractional rate of transport of T3 into tissues is unaltered. For rT3, the production rate is unchanged, but the metabolic clearance rate is decreased, due to decreased extrathyroidal deiodination of rT3 to 3,3′-diiodothyronine. For T4 the metabolic clearance rate is usually normal, but it may be increased in patients with severe illnesses who have low serum T4concentrations, due at least in part to decreased production of one or more serum thyroid hormone–binding proteins. These severely ill patients also tend to have low serum TSH concentrations; therefore, they have low T4production rates. The fractional rate of transport of T4 from serum to tissues is reduced to ~50% of the normal value.

FIGURE 11D.2. Effect of nonthyroidal illness on serum concentrations, metabolic clearance rate (MCR), and production rate (PR) of triiodothyronine (T3) and reverse T3 (rT3; expressed as a percentage of the values in normal subjects). (From Doctor R, Krenning EP, de Jong M, et al. The sick euthyroid syndrome: changes in thyroid hormone serum parameters and hormone metabolism. Clin Endocrinol (Oxf) 1993;39:499, with permission.)

Changes in Hypothalamic–Pituitary–Thyroid Function

The failure of serum TSH concentrations to increase despite marked reductions in serum T3 and sometimes T4 concentrations, as occur in patients with thyroid disease, suggests that the sensitivity of TSH secretion to low serum thyroid hormone concentrations is decreased in patients with nonthyroidal illness. This could be due to changes in the thyrotrophs or to decreased TRH secretion. Evidence for the latter is a decrease in the amplitude (but not frequency) of the nocturnal pulses of TSH secretion; such decreases in pulse amplitude have been documented during fasting (22), after surgery (23), and in a wide variety of nonthyroidal diseases, including chronic renal insufficiency, diabetes mellitus, acute respiratory failure, infections, cancer, and critical illness (24,25,26,27,28). For example, fasting for 60 hours results in disappearance in the nocturnal surge in TSH secretion and a 50% reduction in 24-hour mean serum TSH concentrations (Fig. 11D.3), as well as decreases in serum T3 and increases in serum rT3 concentrations (22). Among patients with nonthyroidal illnesses, the nocturnal surge in TSH secretion is decreased in at least half, and the decrease is not closely related to changes in serum T3 or free T4concentrations (26). The pattern of 24-hour TSH secretion in these patients, therefore, is similar to that in many patients with central hypothyroidism (Fig. 11D.3) (29). If the abnormal pattern of TSH secretion persists for a week or longer in patients with nonthyroidal illness, then their Tproduction rate should decrease, as it does in patients with central hypothyroidism.

FIGURE 11D.3. Serum TSH concentrations measured at frequent intervals for 24 hours in (top) 8 normal subjects before and after fasting for 60 hours, and (bottom) in 21 patients with nonthyroidal illness and 6 patients with central hypothyroidism caused by a pituitary adenoma. (From Romijn JA, Adriaanse R, Brabant G, et al. Pulsatile secretion of thyrotropin during fasting: a decrease of thyrotropin pulse amplitude. J Clin Endocrinol Metab 1990; 71:1631; and Adriaanse R, Romijn JA, Brabant G, et al. Pulsatile thyrotropin secretion in nonthyroidal illness. J Clin Endocrinol Metab 1993;77:1313, with permission.)

In a study of premortem serum thyroid hormone concentrations and postmortem hypothalamic TRH gene expression in patients with nonthyroidal illness, the TRH messenger RNA (mRNA) content in the paraventricular nucleus was positively correlated with serum T3 and TSH concentrations, but not with serum T4 or free T4 concentrations (Fig. 11D.4) (30). In another study, the T3 content of the hypothalamus and anterior pituitary were, respectively, 64% and 46% lower in patients who died as a result of nonthyroidal illness, as compared with patients who died suddenly (31). Furthermore, the TSH that is secreted in patients with nonthyroidal illness is less glycosylated, and therefore less biologically active, than normal, a change that also occurs in patients with overt hypothalamic disease and TRH deficiency (32,33).

FIGURE 11D.4. Top: Macroscopic film autoradiograms of hypothalamic sections of two patients, showing the hybridization signal of thyrotropin-releasing hormone (TRH) messenger RNA (mRNA) in the paraventricular nucleus along the wall of the third ventricle. The left panel shows a low-intensity signal of a patient with low serum triiodothyronine (T3), thyroxine (T4), and thyrotropin (TSH) concentrations measured < 24 hours before death. A high-intensity signal is seen in the right panel of a patient with normal serum concentrations who died from cardiac arrest. Bottom: Relationship between premortem serum TSH or T3 concentrations and total TRH mRNA in the paraventricular nucleus of 10 patients with nonthyroidal illness. To convert serum T3 values to ng/dL, multiply by 65.1. (From Fliers E, Guldenaar SEF, Wiersinga WM, et al. Decreased hypothalamic thyrotropin-releasing hormone gene expression in patients with nonthyroidal illness. J Clin Endocrinol Metab 1997;82:4032, with permission.)

Little is known about the function of the thyroid gland itself in patients with nonthyroidal illness. The increase of serum TSH induced by exogenous TRH or that occurs during recovery from illness is followed by an increase in serum T3 and T4 concentrations, indicating that the thyroid gland can respond normally to TSH. Thyroid weight is lower and thyroid follicular size is reduced in deceased chronically ill patients, as compared with subjects with sudden death (34), possibly related to decreased TSH secretion in severe nonthyroidal illness.

Changes in Peripheral Thyroid Hormone Metabolism

In biopsies taken from intensive-care-unit patients immediately after death, liver iodothyronine deiodinase type 1 (D1) activity was decreased, whereas iodothyronine 5′-deiodinase type 3 (D3) activity was increased in liver and skeletal muscle (see Chapter 7) (35). Their premortem serum T3/rT3 ratio correlated positively with the D1 activity and negatively with the D3 activity in the liver; the levels of D1 and D3 mRNA levels corresponded with the activity of the respective enzyme. The fall in D1 activity, in accord with findings in tissues from starved and sick animals (36), will decrease the extrathyroidal conversion of T4 to T3 and of rT3 to 3,3′-diiodothyronine, thereby contributing to low serum T3 concentrations and high serum rT3 concentrations. The remarkable induction of the activity of D3, which degrades T3 to 3,3′-diiodothyronine, thereby further decreases the availability of T3. D3 is not detectable in normal liver or muscle, but is found in cardiac muscle in animals with heart failure (37).

The changes in iodothyronine deiodination also affect the serum concentrations of diiodothyronines (38,39). Nondeiodinative pathways of iodothyronine degradation are frequently increased in patients with nonthyroidal illness; serum concentrations of T3 sulfate are increased (40,41), although those of T4sulfate are normal (42). There may be increases in the products of ether-link cleavage, such as diiodotyrosine, and in the products of oxidative deamination of the alanine side chain of T4 and T3, which are, respectively, tetraiodothyroacetic acid and triiodothyroacetic acid, in patients with nonthyroidal illness (43).

There are few data on T3-nuclear receptors in patients with nonthyroidal illness (see Chapter 8). Liver T3-nuclear receptor-α1 and -β1 mRNA and protein levels in patients with hepatic diseases were unchanged, as compared with normal liver (44). In contrast, the number of receptors, but not their affinity, was decreased in the livers of animals that were fasted or had diabetes or cancer (7), and lipopolysaccharide injection in mice caused a major fall in T3-nuclear receptor-α1, -α2 and -β1 mRNA and protein expression in liver, and binding of the T3-nuclear receptor/retinoid X receptor heterodimers to DNA was reduced (45). The factors that cause these changes are unknown.

The thyroid hormone content of tissues is decreased in patients with nonthyroidal illness. The mean T3 concentrations in the cerebral cortex, liver, kidney, and lung were lower by 46% to 76% in patients who died of nonthyroidal illness, as compared with those who died suddenly, but the values in heart and skeletal muscle were similar (31). The mean T4 concentration in the liver was 66% lower in the patients with nonthyroidal illness, but the values in the cerebral cortex were similar in the two groups.

Thyroid Hormone Action in Tissue

Most if not all of the changes described earlier would seem to indicate that nonthyroidal illness leads to reduction in thyroid hormone production, especially in the production of T3, and therefore that the availability of T3 in tissues should be reduced. This raises the important but still incompletely answered question of what the level of thyroid hormone action is in the tissues of patients with nonthyroidal illness. During fasting, oxygen consumption and protein breakdown decrease so that energy is saved and organ function may be preserved. The lower T3 production rate during fasting might contribute to this useful metabolic adaptation; similar changes occur in hypothyroidism. Prolongation of the Achilles tendon reflex relaxation time in patients with anorexia nervosa and of the pulse-wave arrival time during hypocaloric feeding are compatible with a hypothyroid-like state in undernutrition (Table 1) (46,47).

TABLE 11D.1. SIMILARITIES AND DISSIMILARITIES BETWEEN CHANGES IN INTERMEDIARY METABOLISM AND PERIPHERAL TISSUE FUNCTION TESTS DURING FASTING, HYPOTHYROIDISM AND ILLNESS


 

Fasting

Hypothyrodism

Illness


Resting energy expenditure

Glucose utilization

Proteolysis

Lipolysis

Fat oxidation

Pulse-wave arrival time (QKd) (ref. 46)

=

Achilles tendon reflex half-relaxation time (ref. 47)

=

Serum angiotensin-converting enzyme (ref. 48)

 ↓

Serum sex hormone-binding globulin (ref. 49)

=

Serum Osteocalcin (ref. 49)

 ↓

=

Erythrocyte Na+, K+-ATPase (ref. 50)

 ↑


↑, increase; ↓decrease; =, no change; Na+, K+-ATPase, sodium, potassium-adenosine triphosphatase; QKd, time interval between Q-wave electrocardiogram and arrival of pulse wave in brachial artery.

In contrast, the illness-induced changes in glucose, protein, and fat metabolism are opposite to those that occur in hypothyroidism (48,49,50). In theory, then, the occurrence of the low T3 syndrome could be a beneficial adaptation to illness, serving to minimize its catabolic effects.

PATHOGENESIS OF CHANGES IN THYROID FUNCTION IN NONTHYROIDAL ILLNESS

Inhibition of Iodothyronine 5′-Deiodinase

The decreased rate of extrathyroidal production of T3 and decreased metabolic clearance rate of rT3 during fasting and illness can be explained by a decrease in activity of D1, which catalyses both the 5′-monodeiodination of T4 to T3 and of rT3 to 3,3′-diiodothyronine. The decrease in D1 activity may be due to inhibition by drugs such as high doses of glucocorticoids and propranolol or amiodarone. The increase of cortisol secretion during illness does not appear to contribute to the decrease in D1 activity, because there is no consistent relationship between serum cortisol and T3 concentrations in patients with nonthyroidal illness, and the decrease in serum T3 concentrations after abdominal hysterectomy persists despite abolition of the postoperative increase in serum cortisol by afferent nerve blockade (51).

The decrease in liver D1 activity of patients with critical illness (35) might be causally related to cytokines, in view of the capacity of interleukin (IL)-1β, IL-6, tumor necrosis factor (TNF)-α, and interferon-γ to decrease D1 mRNA in vitro (52). Indeed, IL-1 and IL-6 inhibit the T3-induced induction of D1 mRNA and enzyme acitivity in cultured rat hepatocytes; this transcriptional effect can be partially overcome by exogenous steroid receptor coactivator-1 (SRC-1) (53). The proposed mechanism is competition for limiting amounts of nuclear receptor coactivators between the D1 promotor and the promoters of cytokine-induced genes. TNF-α impairs T3-dependent induction of D1 gene expression by activation of nuclear factor-kappa B, (NF-kB), possibly via sequestration of common cofactors shared between NF-kB and thyroid hormone receptors (54).

Inhibition of Plasma Membrane Transport of Iodothyronines

T4 and T3 enter cells by active, energy-dependent transport across the plasma membrane of the cells. Several biologic compounds inhibit cell uptake of T4 and T3, as determined by incubation of cultured rat hepatocytes with radioiodine-labeled T4 or T3 and subsequent measurements of iodide and T4 or T3 conjugates in the medium. This system allows discrimination between inhibition of transport (lowered iodide production without accumulation of conjugates) and inhibition of deiodination (lowered iodide production with accumulation of sulfates and glucuronides). In these cultures, the furan fatty acid 3-carboxy-4-methyl-5-propyl-2-furan propanoic acid (CMPF), indoxyl sulfate, and hippuric acid all inhibit iodide production, and conjugates do not accumulate. The serum of patients with chronic renal insufficiency has similar effects; the serum concentrations of CMPF and indoxyl sulfate (but not of hippuric acid) in these patients are above the minimum concentration required for in vitro inhibition of iodide production by these compounds (55). It follows that CMPF and indoxyl sulfate may be responsible for the low T3 syndrome in patients with chronic renal insufficiency by inhibiting cellular uptake and subsequent deiodination of T4. These compounds do not affect the uptake of T4 or T3 by cultured rat anterior pituitary cells, nor do they interfere with TSH release (56). Thus, the uptake of thyroid hormones into the pituitary is regulated differently than uptake into the liver.

Bilirubin and nonesterified fatty acids like oleic acid also inhibit T4 transport into hepatocytes. Inhibition of T4 transport by the serum of critically ill patients with high serum bilirubin concentrations correlated with the molar ratio of bilirubin to albumin and that of nonesterified fatty acids to albumin (57). There may be other inhibitors of T4 transport into cells as well. In a study of serum samples from patients with mild, moderate, or severe nonthyroidal illness, the degree of inhibition of T4 transport into rat hepatocytes was inversely correlated with the patients' serum T3 concentrations (58).

The liver is a major site for production of T3 and clearance of rT3, and inhibition of transport of T4 and rT3 (which share the same transport mechanism) into the liver might thus contribute to the characteristic decreases in production of T3 and clearance of rT3 that occur in patients with nonthyroidal illness. This fits with data indicating that cellular uptake of thyroid hormones is rate limiting for subsequent intracellular metabolism and nuclear binding, and with the observed reduction of the fractional transfer rate of T4 from plasma to tissues in patients with nonthyroidal illness (59). Further support for this hypothesis comes from studies in human subjects in whom the low T3 syndrome caused by caloric deprivation was associated with decreased T3 and especially T4 uptake into the liver (60,61). The decrease in uptake, which is an energy-dependent process, could well be due to a low adenosine triphosphate (ATP) content in the liver, as has been demonstrated in starvation and nonthyroidal illness by magnetic resonance spectroscopy (62). Fructose, which lowers hepatic ATP content, indeed decreases hepatic T4 uptake (63).

Inhibition of Thyroxine Binding to Serum Proteins

Serum concentrations of thyroxine-binding globulin (TBG), the main carrier protein for T4, are usually normal in patients with nonthyroidal illness, but some patients have slightly low concentrations. The serum concentrations of the two other carrier proteins, transthyretin and albumin, can be low as well. These decreases result in lower serum total T4 (and T3) concentrations but should not affect serum free T4 concentrations under equilibrium conditions. The rapid decrease of TBG (a member of the serpin superfamily) during sepsis and cardiopulmonary bypass might be due to cleavage by inflammatory serine proteases; the smaller TBG cleavage product has a reduced affinity for T4(64,65).

Another cause of decreased serum binding of T4 is the presence of inhibitors of binding. Both protein and nonprotein inhibitors of T4 binding to TBG or the other thyroid hormone–binding proteins have been found in the serum of patients with nonthyroidal illness in different studies (15,16,66,67,68). The non-protein-binding inhibitors include furosemide, fenclofenac, and salicylate. High doses of these drugs are competitive inhibitors of T4 binding to TBG. They therefore acutely raise serum free T4 concentrations, which inhibits TSH secretion, so that serum free T4 concentrations return to normal; serum T4concentrations decrease because the protein-binding sites are occupied by the drug (see Chapter 13).

Nonesterified fatty acids also have been proposed as a nonprotein inhibitor of T4 binding. Addition of nonesterified fatty acids to normal human serum increases the free T4 fraction by inhibiting T4 binding to albumin, but only if the molar ratio of nonesterified fatty acids to albumin is higher than 5:1, a value rarely reached even in critically ill patients (69). Heparin is a weak inhibitor of T4 binding to TBG. Its action is indirect, via activating serum lipoprotein lipase that then breaks down triglycerides to nonesterified fatty acids (and glycerol) in vitro, and can be prevented by adding protamine to the serum in vitro (70). Thus, these substances are unlikely to affect T4 binding appreciably in vivo.

Inhibitory Effects of Cytokines

Cytokines are important mediators of the acute phase response to tissue injury. It is thus plausible to suppose that cytokines cause some of the changes in hypothalamic–pituitary–thyroid function that occur in patients with nonthyroidal illness. Indeed, some cytokines, such as IL-1, IL-6, TNF-α, and interferon-γ inhibit the synthesis or secretion of TSH, thyroglobulin, T3, and thyroid hormone–binding proteins and decrease D1 mRNA and the binding capacity of T3-nuclear receptors in vitro in cultured human and animal cells (52,71). IL-1 and TNF-α act mainly in a paracrine and autocrine fashion; their serum concentrations are mostly undetectable or low, even during illness, and no quantitative relationships have been found between serum T3 or T4 concentrations and the serum concentrations of either of these cytokines in patients with nonthyroidal illness (72,73,74). Similarly, serum T3 concentrations are not correlated with serum interferon-γ, IL-8, and IL-10 concentrations in these patients (75). In contrast, IL-6 is a systemic cytokine that has endocrine actions, and there is an inverse correlation between serum T3 and IL-6 concentrations in patients with nonthyroidal illness (Fig. 11D.5) (74,76,77,78).

FIGURE 11D.5. The relationship between serum triiodothyronine (T3) and serum interleukin (IL)-6 concentrations (top) and a cytokine score, derived from the summation of serum IL-6, soluble tumour necrosis factors (TNF) receptors (sTNFaRp55 and sTNFaRp75), soluble IL-2 receptor, and IL-1 receptor antagonist concentrations in individual serum samples (bottom) in consecutive hospitalized patients with nonthyroidal disease. To convert serum T3values to ng/dL, multiply by 65.1. (From Boelen A, Platvoet-ter Schiphorst M, Wiersinga WM. Association between serum interleukin-6 and serum 3,5,3′-triiodothyronine in nonthyroidal illness. J Clin Endocrinol Metab 1993;77:1695; and Boelen A, Platvoet-ter Schiphorst MC, Wiersinga WM. Soluble cytokine receptors and the low 3,5,3′-triiodothyronine syndrome in patients with nonthyroidal disease. J Clin Endocrinol Metab 1995;80:971, with permission.)

Soluble forms of membrane-bound cytokine receptors or receptor antagonists have the ability to modulate cytokine activity and are generated in response to the same inflammatory stimuli that induce cytokine production, and the serum concentrations of these proteins reflect the degree of activation of the cytokine network. Serum T3 concentrations are negatively correlated with the serum concentrations of soluble

TNF receptors (sTNFaRp55 and sTNFaRp75), soluble IL-2 receptors, and IL-1 receptor antagonist in patients with nonthyroidal illness (79). A cytokine score composed of the serum concentrations of soluble cytokine receptors and IL-6 is strongly inversely correlated with serum T3 concentrations in these patients (Fig. 11D.5). The serum concentrations of C-reactive protein, an acute-phase protein induced by IL-6, also are inversely correlated with serum T3concentrations in such patients (74,78).

Administration of single doses of TNF-α, interferon-α, or IL-6 to normal subjects results within hours in a 12% to 25% decrease in serum T3 concentrations, a 36% to 72% increase in serum rT3 concentrations, and a 13% to 32% decrease in serum TSH concentrations, whereas serum total and free T4 concentrations change little (80,81,82). A single dose of interferon-γ had no effect (83). Despite the resemblance of these changes to those that occur in patients with nonthyroidal illness, the findings do not prove that cytokines cause the changes in patients with nonthyroidal illness. Furthermore, administration of cytokines causes an influenza-like illness, with headache, nausea, and fever, which may cause the hormonal changes via other mechanisms. Chronic administration of cytokines to humans or experimental animals also results in transient changes in thyroid hormone metabolism and regulation, possibly explained by the capacity of cytokines—which operate both as a cascade and as a network—to modulate the production of other cytokines and cytokine receptors.

Administration of bacterial endotoxin to normal subjects is followed by a fall in serum T3, free T4, and TSH concentrations, and a rise in serum rT3concentrations. Blocking the action of IL-1 by coadministration of IL-1 receptor antagonist or the action of TNF-α by administration of recombinant dimeric TNF receptors has no effect on the endotoxin-induced changes (84,85). The results of studies of immunoneutralization of cytokines in animals have also been inconclusive (86). The best evidence to date for a causal role of cytokines in the pathogenesis of the changes in nonthyroidal illness comes from experiments with IL-6 knock-out mice (87), in which the decrease in serum T3 concentrations during induced illness is less than that of normal mice; hepatic D1 activity does not change in the knock-out mice but decreases in normal mice.

Inhibition of Thyrotropin-Releasing Hormone and Thyrotropin Secretion

The changes in the relationship between serum TSH and serum thyroid hormone concentrations in patients with nonthyroidal illness indicate a decrease in the sensitivity of TSH secretion to decreases in serum T4 and T3 concentrations. A decrease in sensitivity is further suggested by the smaller than normal increase in serum TSH concentrations in response to the small decreases in serum T4 and T3 concentrations induced by administration of inorganic iodide that occurs in patients with nonthyroidal illness (88). The cause of the change in sensitivity of the thyrotroph cells is not known, but the decrease in TRH gene expression in the paraventricular nucleus of the hypothalamus in patients with nonthyroidal illness is correlated with the decrease in serum TSH concentrations (30). Changes in brain thyroid-hormone metabolism or in serum cortisol and leptin concentrations, as well as induction of cytokines, could be involved in the down-regulation of TRH (see section on regulation of thyrotropin secretion in Chapter 10).

In humans, T3-nuclear receptors are found in the hypothalamic paraventricular nucleus and infundibular nucleus (the human homologue of the rat arcuate nucleus). Iodothyronine-5′-deiodinase type 2 (D2) is—at least in rats—not expressed in the paraventricular nucleus, but is present in the arcuate nucleus in tanycytes, specialized ependymal cells lining the wall of the third ventricle near the median eminence of the hypothalamus. In fasted rats, D2 activity in the arcuate nucleus is upregulated: the increase of locally produced T3 is likely responsible for lower TRH synthesis in the paraventricular nucleus, but which neurons of the arcuate nucleus projecting to the paraventricular nucleus mediate this effect is unknown (89). In systemic illness induced by bacterial lipopolysaccharide in rats, D2 activity increased markedly in the anterior pituitary and in tanycytes; the locally generated T3 may suppress the synthesis of pituitary TSH as well as that of TRH (90).

Adrenalectomy in rats results in an increase in TRH mRNA in the paraventricular nucleus, and administration of glucocorticoid reverses the increase (91). However, the hypercortisolism of critically ill patients probably does not contribute to the fall in TRH production, because TRH production remains low when adrenalectonized animals treated with corticosterone are given lipopolysaccharide (92).

The adipocyte-derived hormone leptin has been recognized as an important mediator of neuroendocrine changes during illness. Leptin binds to leptin receptors in the infundibular nucleus where neuropeptide-Y (NPY)-containing cells are key targets. NPY, agouti-related protein (AGRP), and α-melanocyte stimulating hormone (α-MSH) neurons project directly from the infundibular nucleus to hypophysiotropic TRH neurons in the paraventricular nucleus (93). NPY and AGRP (both suppressed by leptin) inhibit and α-MSH (induced by leptin) stimulates TRH gene expression (83,94,95). In starved animals, the fall in serum leptin concentrations is associated with a decrease of TRH mRNA in the paraventricular nucleus, and administration of leptin reverses the decrease in TRH mRNA (96). This effect is mediated via the arcuate nucleus, because ablation of the arcuate nucleus prevents the decrease in hypothalamic TRH and serum T4concentrations during fasting and prevents the reversal of these changes by leptin (97). In humans, however, administration of leptin only partially prevents the starvation-induced decrease in serum TSH concentrations, and does not prevent the fall in serum T3 concentrations or the rise in rT3 concentrations (98).

In patients with systemic illness serum leptin concentrations are often increased, like they do acutely in response to administration of TNF-α or IL-1, but the increase is not related to serum T3 and T4 concentrations (99,100). In rats, endotoxin increases serum leptin concentrations and propiomelanocortin but not NPY gene expression in the arcuate nucleus, which together would increase rather than decrease hypothalamic TRH production. Furthermore, serum leptin is negatively correlated with NPY mRNA expression in the infundibular nucleus of critically ill patients, and the NPY expression is—in contrast to the situation in food-deprived rodents—positively correlated with TRH mRNA expression in the paraventricular nucleus (101). There is thus no evidence that leptin is involved in the changes in peripheral thyroid hormone metabolism; leptin may be partially responsible for the decrease in TSH secretion during fasting, but probably has no role in systemic illness.

The role of cytokines (especially IL-1β) in the activation of the hypothalamic–pituitary–adrenal axis is well known. Cytokines also affect TRH production, at least in rats. The hypothalamic content of TRH mRNA decreases after administration of IL-1, inappropriate for the decrease in serum T3 and T4 concentrations that occurs at the same time (102). Interleukin-1β decreases the release of TSH (but not of other anterior pituitary hormones) in cultured rat anterior pituitary cells, an effect not mediated by thyroid hormone uptake or T3 nuclear receptor occupancy (103). The effect of TNF-α on TSH release is disputed. Interleukin-6 decreases TSH secretion, possibly by a direct effect on the thyrotrophs.

There are few studies on the handling of thyroid hormones in the pituitary during nonthyroidal illness, but the available data do not indicate that thyroid hormone uptake is altered (56). Up-regulation of D2 activity in the mediobasal hypothalamus and anterior pituitary, as found during critical illness, may contribute to suppression of production of TRH and TSH, respectively, and it is tempting to speculate that cytokines are involved. Experimental data support the notion that cytokines induce D2 activity in tanycytes via NF-kB binding sites in the promoter region of the D2 gene (54).

The down-regulation at the hypothalamus–pituitary level provides an explanation for the decreased sensitivity of TSH secretion to low serum T3 and T4concentrations in patients with nonthyroidal illness. Specifically, decreased TRH secretion is the likely cause of the decreased nocturnal TSH surge and the release of TSH of reduced biologic activity, phenomena akin to the changes in some patients with central hypothyroidism.

The changes in hypothalamic–pituitary–thyroid function that occur in patients with nonthyroidal illness thus occur as a result of changes in both the central and peripheral compartments of the thyroid endocrine system (Fig. 11D.6). Although a persistent decrease in TRH and TSH secretion might lead to a decrease in T4 production, the peripheral changes develop independent of changes in the hypothalamus and pituitary. It follows that the factors mediating the various changes are operative simultaneously both centrally and peripherally.

FIGURE 11D.6. A schematic diagram of the hypothalamic–pituitary–thyroid axis showing the changes that occur in patients with nonthyroidal illness (NTI) and the causes of those changes (for explanation, see text). bil, bilirubin; D1, iodothyronine 5′-deiodinase type 1; D3, iodothyronine 5-deiodinase type 3; GSH, glutathione; IS, indoxyl sulfate; MCR, metabolic clearance rate; NEFA, nonesterified fatty acids; PR, production rate; Se, selenium; TR, thyroid receptor; T3S, triiodothyronine (T3) sulfate.

CLINICAL IMPORTANCE OF THE CHANGES IN THYROID FUNCTION IN NONTHYROIDAL ILLNESS

Prognostic Importance

The magnitude of the changes that occur in patients with nonthyroidal illness in general reflects the severity of the illness, and several of the individual changes have been linked to outcome in different studies. In general, the prognosis is poorer in patients with lower serum T3, T4, or TSH concentrations or higher serum rT3 concentrations. In a study of critically ill patients admitted to an intensive care unit, the sensitivity and specificity in predicting mortality were 75% and 80%, respectively, for a serum T4 concentration of < 3.1 µg/dL (40 nmol/L), 56% and 100% for an 8 AM serum cortisol concentration of >30 µg/dL (830 nmol/L), and 100% and 82% for the combination (104).

Diagnosis of Thyroid Disease in Nonthyroidal Illness

The prevalence of nonthyroidal illness depends on what groups of patients are evaluated. Among patients admitted to hospital medical services, the prevalence of a low serum T3 concentration is ~50%, that of a low serum T4 concentration is ~15% to 20%, and that of an abnormal (low or high) serum TSH concentration ~10%. The changes are less frequent among outpatients and more frequent among patients admitted to intensive care units.

The high prevalence of changes among hospitalized patients hampers the diagnosis of thyrotoxicosis or hypothyroidism among them, and therefore argues against screening them for thyroid dysfunction. In patients clinically suspected to have thyroid disease, the single most useful test is measurement of serum TSH. A normal result virtually excludes thyrotoxicosis or hypothyroidism (as it does among outpatients) (105,106).

A high serum TSH concentration is compatible with primary hypothyroidism, but also with nonthyroidal illness, if the elevation is slight (Fig. 11D.7). The combination of high serum TSH and low T4 concentrations increases the likelihood of hypothyroidism, but also can be found in patients recovering from a nonthyroidal illness. The finding of a high ratio of T3 to T4 in serum, a low thyroid hormone–binding ratio, and a low serum rT3 concentration favor the presence of hypothyroidism, because the opposite changes occur in nonthyroidal illness, but the diagnostic accuracy of any of these measurements is limited (107). Furthermore, in patients with hypothyroidism the high serum TSH concentrations may decrease, even into the normal range, during an acute illness, especially in those patients given dopamine or high doses of glucocorticoids (108).

FIGURE 11D.7. The proportion of patients with thyroid disease (Thyr. or Thyroid), nonthyroidal illness (NTI), and receiving glucocorticoid therapy (GC) among a large group of hospitalized patients, subdivided according to five ranges of serum thyrotropin TSH concentrations. (From Stockigt JR. Guidelines for diagnosis and monitoring of thyroid disease: nonthyroidal illness. Clin Chem 1996;42:188, modified with permission; and Spencer CA. Clinical utility and cost-effectiveness of sensitive thyrotropin assays in ambulatory and hospitalized patients. Mayo Clin Proc 1988;63:1214, with permission.)

A low serum TSH value in any patient raises the possibility of thyrotoxicosis, especially if the serum free T4 concentration is high. This combination of test results is unfortunately also found in patients with nonthyroidal illness. However, the greater the extent of TSH suppression, the greater the likelihood of thyrotoxicosis (Fig. 11D.7). The combination of thyrotoxicosis and nonthyroidal illness may cause T4-thyrotoxicosis because the illness causes a decrease in serum T3 concentrations in patients with thyrotoxicosis just as it does in any other patient (Fig. 11D.8).

FIGURE 11D.8. Serum thyroxine (T4), triiodothyronine (T3), and reverse T3 (rT3) concentrations in a patient with thyrotoxicosis hospitalized for treatment of pneumonia. The initial values represented the biochemical pattern of T4-thyrotoxicosis, but with recovery from the pneumonia they evolved to those of classical thyrotoxicosis with high serum T4 and T3 concentrations. Serum thyrotropin values were undetectable at all times. To convert serum T4 values to µg/dL, multiply by 0.078; and to convert serum T3 and rT3 values to ng/dL, multiply by 65.1.

In some of these situations, history, physical examination, and the presence of high serum thyroid antibody concentrations can provide further clues to the presence or absence of thyroid disease. In many patients, however, no definite diagnosis can be established, and the most prudent policy is to repeat the thyroid function tests after the nonthyroidal illness has improved.

Thyroid Hormone Treatment of Patients with Nonthyroidal Illness

In obese subjects starting a hypocaloric diet, maintenance of baseline serum T3 concentrations by administration of T3 prevents the occurrence of hypothyroid-like effects such as a decrease in pulse rate and oxygen consumption and prolongation of the pulse-wave arrival time (46,109). In fasting normal subjects, prevention of the decrease in serum T3 concentrations by exogenous T3 administration increases protein breakdown and inhibits TSH secretion (110). These results suggest that the occurrence of the low T3 syndrome during periods of inadequate caloric intake has protective value in conserving energy and limiting proteolysis. However, administration of T3 during more prolonged fasting does not increase proteolysis, but does increase hepatic glucose appearance (109,111,112). The role of insulin might be important in this respect, because decreasing insulin secretion during fasting is a critical determinant of the shift from carbohydrate- to fat-based metabolism.

In severely ill patients with serum T4 concentrations < 5.0 µg/dL (65 nmol/L), intravenous administration of replacement doses of T4 (1.5 µg/kg/day) for 2 weeks did not reduce mortality as compared with untreated patients; during treatment the patients given T4 had serum T3 concentrations similar to those in the other group, but their serum T4 concentrations were higher and their serum TSH concentrations were lower (113). Likewise, administration of T3 to patients with severe burns did not reduce mortality despite restoration of normal serum free T3 index values (114), and T4 given intravenously to patients with acute renal failure actually increased mortality (115). In patients with chronic renal insufficiency, T3 administration increased proteolysis (116). Finally, administration of T4 or T3 to animals with nonthyroidal illness has no clinically beneficial effect and may in fact increase mortality (117,118,119). Taken together, the results of these studies argue against thyroid-hormone treatment of patients with nonthyroidal illness, including those with low serum TSH and T4 concentrations, and instead suggest that the changes are beneficial adaptations to illness.

In general, the acute-phase response to illness is not uniformly beneficial: when extreme, it can be fatal, as for example in patients with septic shock. The possibility remains that the changes may represent maladaptation. In adult patients undergoing coronary artery bypass surgery, serum T3 concentrations decrease by ~40% within minutes or hours after the operation. In studies in which supraphysiologic doses of T3 were given starting immediately after completion of bypass surgery, resulting in supranormal serum T3 concentrations, the cardiac index was slightly higher and systemic vascular resistance was slightly lower in the patients given T3, as compared with placebo. The need for inotropic drugs or mechanical assistance, the incidence of atrial fibrillation, and postoperative morbidity and mortality were lower in some but not all studies (120,121,122,123). In children with congenital heart disease T3 treatment, as compared with placebo, after cardiac surgery also increased cardiac output and decreased systemic vascular resistance (124,125). The overall utility of T3administration after cardiac surgery is still not fully established.

TABLE 11D.2. CHANGES IN SERUM THYROID HORMONE AND THYROTROPIN CONCENTRATIONS IN PTIENTS WITH NON THYROIDAL ILLNESS


 

Serum T3

Serum rT3

Serum T4

Serum free T4

Serum TSH


Fasting

=

=

Mild illness

=

=, ↑

=

Critical illness

=, ↓

Surgical trauma, burns

=, ↓

Chronic renal failure

=

=, ↓

=, ↓

=, ↓

Hepatitis

=, ↑

=, ↑

=, ↑

=, ↓

=

HIV infection

=

=

=, ↓

=, ↑

Depression

=, ↓

=, ↑

=, ↑

=, ↑

=, ↓


=, no change; ↓ or ↑; direction of change compared with values in normal subjects; HIV, human immunodeficiency virus; rT3, reverse triiodothyonine; T4 thyroxine; TSH, thyrotropin.

Lastly, co-infusion of growth hormone–releasing peptide (GHRP-2) plus TRH (but not TRH alone) in protracted critical illness induced restoration of pulsatile growth hormone and TSH release, near-normal serum T4 and T3 values, and a shift toward anabolic metabolism (126). This interesting finding corroborates the neuroendocrine component of nonthyroidal illness.

SPECIFIC TYPES OF NONTHYROIDAL ILLNESS

Chronic Renal Insufficiency

Serum T3 concentrations are low, but serum rT3 concentrations may not be high in patients with chronic renal insufficiency (Table 11D.1). This poses a conceptual problem because hepatic uptake of rT3 as well as that of T4 is thought to be inhibited in these patients. Serum free T3 and free T4 concentrations are lower in the patients than in normal subjects, independent of the assay method used, and increase slightly after hemodialysis (126,127). After transplantation, serum T3 concentrations reflect graft function (128). A high frequency of goiter, up to 43% in one study, has been described in patients with chronic renal insufficiency, and has been attributed to high serum inorganic iodide concentrations (129).

Liver Disease

Patients with cirrhosis have changes in serum thyroid hormone and TSH concentrations similar to those of patients with nonthyroidal illness in general. In those with infectious hepatitis, serum T4 and T3 concentrations are often high, due to an increase in serum TBG concentrations; their serum free T4 and T3 concentrations are usually normal. The findings are similar in patients with chronic active hepatitis and primary biliary cirrhosis, except in those with substantial hepatic insufficiency, in whom the concentrations decrease.

Human Immunodeficiency Virus Infection

Patients with asymptomatic human immunodeficiency virus (HIV) infection or those with the acquired immunodeficiency syndrome (AIDS) who have no opportunistic infections or hepatic dysfunction have normal serum T4 and T3 concentrations. Their serum free T4 index values and free T4 concentrations are normal or slightly low (130,131). Some of the patients have slightly high serum TBG concentrations, which tend to be inversely related to the percentage of CD4 cells, and serum TBG values have been used as a surrogate marker for disease progression (132). Serum TSH concentrations can be slightly high as well, but the small changes in TBG and TSH are different from those in nonthyroidal illness, and their pathogenesis and relevance are unknown. The prevalence of subclinical hypothyroidism may be increased in patients with HIV infection, especially in patients receiving highly active antiretroviral therapy (HAART) (133,134,135). The HAART-associated lipodystrophy characterized by peripheral fat wasting and central adiposity is apparently not related to thyroid function (136).

Patients with AIDS complicated by Pneumocystis carinii infections or other serious infections have the typical findings of patients with other severe nonthyroidal illness (137).

Psychiatric Disease

Patients with posttraumatic stress disorder may have slightly high serum T4, T3, and TBG concentrations, but their serum free T4 and free T3 concentrations are normal (138). Among patients with acute psychosis who are hospitalized, ~10% have some abnormality in thyroid function, including high or low serum TSH concentrations and high serum T4 concentrations (139,140). The changes are somewhat dependent on the particular psychiatric disorder; patients with substance abuse are more likely to have high serum TSH concentrations, whereas those with mood disorders tend to have high serum T4 concentrations (with low, normal, or high serum TSH concentrations). The tests usually become normal in 7 to 10 days.

Patients with rapid cycling bipolar disorder may have slightly low serum T4 concentrations and high serum TSH concentrations. Those with major depression tend to have higher than average serum T4 and free T4 concentrations (although the values are usually within the respective normal ranges) and low basal and TRH-stimulated serum TSH concentrations (141), in line with a decreased TRH mRNA in the paraventricular nucleus (142). The hypercortisolism and hyperthyroxinemia in patients with depression may inhibit D2 activity in the brain, resulting in low intracerebral T3 content. This may exaggerate the decrease in brain serotonin concentrations that occurs in depressed patients. Such a mechanism may explain why administration of T3 might be effective in patients with depression refractory to tricyclic antidepressant drugs, as has been reported in some but not all studies (143,144). A recent study, however, found that the efficacy and the rapidity of the response to selective serotonin reuptake inhibitors was not increased by adding T3 (145).

CONCLUSION

Patients with nonthyroidal illness have many changes in hypothalamic–pituitary–thyroid function, the magnitude of which roughly correlates with the severity of the illness. The changes include central down-regulation by inhibition of the hypothalamus–pituitary–thyroid axis and peripheral down-regulation by inhibition of the extrathyroidal production of T3. The central and peripheral changes can occur independently of each other. The pathogenesis of these changes is almost certainly multifactorial, with various mechanisms simultaneously operative at different levels. Overall, the changes are apparently geared to limit the impact of T3 on its target organs, which from a teleologic point of view might be a beneficial adaptation during starvation and illness to conserve energy and substrates and minimize excessive tissue catabolism. Although severe nonthyroidal illness may represent maladaptation, the limited data available suggest that administration of thyroid hormone to patients with nonthyroidal illness does not improve the outcome of their illness, but high doses of T3 given directly after cardiac surgery may have some clinical utility. The ill-chosen term “nonthyroidal illness syndrome” might better be replaced by thyroid hormone adaptation syndrome.

REFERENCES

1. Van den Berghe G, de Zegher F, Bouillon R. Acute and prolonged critical illness as different neuroendocrine paradigms. J Clin Endocrinol Metab 1998;83:1827.

2. Chopra IJ. Euthyroid sick syndrome: is it a misnomer? J Clin Endocrinol Metab 1997;82:329.

3. DeGroot LJ. Dangerous dogmas in medicine: the nonthyroidal illness syndrome. J Clin Endocrinol Metab 1999;84:151.

4. Wartofsky L, Burman KD. Alterations in thyroid function in patients with systemic illness: the “euthyroid sick syndrome.” Endocr Rev 1982;3:164.

5. Wartofsky L. The low T3 or “sick euthyroid syndrome”: update 1994. Endocr Rev 1994;3:248.

6. Docter R, Krenning EP, de Jong M, et al. The sick euthyroid syndrome: changes in thyroid hormone serum parameters and hormone metabolism. Clin Endocrinol (Oxf) 1993;39:499.

7. Tibaldi JM, Surks MI. Animal models of nonthyroidal disease. Endocr Rev 1985;6:87.

8. Mak YT, Chan ELP, Chan A, et al. Free triiodothyronine in sera of acutely ill general medical patients: a prognostic indicator? Clin Chem 1992;38:414.

9. Sapin R, Schlienger JL, Kaltenbach G, et al. Determination of free triiodothyronine by six different methods in patients with non-thyroidal illness and in patients treated with amiodarone. Ann Clin Biochem 1995;32:314.

10. Faber J, Siersbaek-Nielsen K. Serum free 3,5,3′-triiodothyronine (T3) in non-thyroidal somatic illness, as measured by ultrafiltration and immunoextraction. Clin Chim Acta 1996; 256:115.

11. Chopra IJ. Simultaneous measurement of free thyroxine and free 3,5,3′-triiodothyronine in undiluted serum by direct equilibrium dialysis/radioimmunoassay: evidence that free triiodothyronine and free thyroxine are normal in many patients with the low triiodothyronine syndrome. Thyroid 1998; 8:249.

12. Surks MI, Hupart KH, Pan C, et al. Normal free thyroxine in critical nonthyroidal illnesses measured by ultrafiltration of undiluted serum and equilibrium dialysis. J Clin Endocrinol Metab 1988;67:1031.

13. Wong TK, Pekary AE, Hoo GS, et al. Comparison of methods for measuring free thyroxine in nonthyroidal illness. Clin Chem 1992;38:720.

14. Docter R, Toor H van, Krenning EP, et al. Free thyroxine assessed with three assays in sera of patients with nonthyroidal illness and of subjects with abnormal concentrations of thyroxine-binding proteins. Clin Chem 1993;39:1668.

15. Chopra IJ, Teco GNC, Nguyen AH, et al. In search of an inhibitor in non-thyroidal illnesses. J Clin Endocrinol Metab 1979;49:63.

16. Oppenheimer JH, Schwartz HL, Mariash CN, et al. Evidence for a factor in the sera of patients with non-thyroidal disease which inhibits iodothyronine binding to solid matrices, serum proteins, and rat hepatocytes. J Clin Endocrinol Metab 1982; 54:757.

17. Hamblin PS, Dyer SA, Mohr VS, et al. Relationship between thyrotropin and thyroxine changes during recovery from severe hypothyroxinemia of critical illness. J Clin Endocrinol Metab 1986;62:717.

18. Feelders RA, Swaak AJG, Romijn JA, et al. Characteristics of recovery from the euthyroid sick syndrome induced by tumor necrosis factor alpha in cancer patients. Metabolism 1999;48: 324.

19. Wehman RE, Gregerman RI, Burns WH, et al. Suppression of thyrotropin in the low-thyroxine state of severe nonthyroidal illness. N Engl J Med 1985;312:546.

20. Sumita S, Ujike Y, Namiki A, et al. Suppression of the thyrotropin response to thyrotropin-releasing hormone and its association with severity of critical illness. Crit Care Med 1994; 22:1603.

21. Kaptein EM. Thyroid hormone metabolism in illness. In: Hennemann G, ed. Thyroid hormone metabolism. New York: Marcel Dekker Inc, 1986:297.

22. Romijn JA, Adriaanse R, Brabant G, et al. Pulsatile secretion of thyrotropin during fasting: a decrease of thyrotropin pulse amplitude. J Clin Endocrinol Metab 1990;71:1631.

23. Bartalena L, Martino E, Brandi LS, et al. Lack of nocturnal serum thyrotropin surge after surgery. J Clin Endocrinol Metab 1990;70:293.

24. Wheatly T, Clark PMS, Clark JDA, et al. Abnormalities of thyrotropin (TSH) evening rise and pulsatile release in haemodialysis patients: evidence for hypothalamic-pituitary changes in chronic renal failure. Clin Endocrinol (Oxf) 1989;31:39.

25. Arem R, Deppe S. Fatal nonthyroidal illness may impair nocturnal thyrotropin levels. Am J Med 1990;88:258.

26. Romijn JA, Wiersinga WM. Decreased nocturnal surge of thyrotropin in nonthyroidal illness. J Clin Endocrinol Metab 1990; 70:35.

27. Bartalena L, Cossu E, Grasso L, et al. Relationship between nocturnal serum thyrotropin peak and metabolic control in diabetic patients. J Clin Endocrinol Metab 1993;76:983.

28. Adriaanse R, Romijn JA, Brabant G, et al. Pulsatile thyrotropin secretion in nonthyroidal illness. J Clin Endocrinol Metab 1993; 77:1313.

29. Adriaanse R, Brabant G, Endert E, et al. Pulsatile thyrotropin release in patients with untreated pituitary disease. J Clin Endocrinol Metab 1993;77:205.

30. Fliers E, Guldenaar SEF, Wiersinga WM, et al. Decreased hypothalamic thyrotropin-releasing hormone gene expression in patients with nonthyroidal illness. J Clin Endocrinol Metab 1997;82:4032.

31. Arem R, Wiener GG, Kaplan SG, et al. Reduced tissue thyroid hormone levels in fatal illness. Metabolism 1993;42: 1102.

32. Lee H-Y, Suhl J, Pekary AE, et al. Secretion of thyrotropin with reduced concanavalin-A binding activity in patients with severe nonthyroidal illness. J Clin Endocrinol Metab 1987;65: 942.

33. Magner J, Roy P, Fainter L, et al. Transiently decreased sialylation of thyrotropin (TSH) in a patient with the euthyroid sick syndrome. Thyroid 1997;7:55.

34. DeJongh FE, Jöbsis AC, Elte JWF. Thyroid morphology in lethal non-thyroidal illness: a post-mortem study. Eur J Endocrinol 2000;144:221.

35. Peeters RP, Wouters PJ, Kaptein E, et al. Reduced activitation and increased inactivation of thyroid hormone in tissues of critically ill patients. J Clin Endocrinol Metab 2003;88: 3202.

36. Bianco AC, Salvatore D, Gereben B, et al. Biochemistry, cellular and molecular biology, and physiological roles of the iodothyronine selenodeiodinases. Endocr Rev 2002;23:38.

37. Wassen FW, Schiel AE, Kuiper GG, et al. Induction of thyroid hormone-degrading deiodinase in cardiac hypertrophy and failure. Endocrinology 2002;143:2812.

38. Pinna G, Hiedra L, Meinhold H, et al. 3,3′-diiodothyronine concentrations in the sera of patients with nonthyroidal illnesses and brain tumors and of healthy subjects during acute stress. J Clin Endocrinol Metab 1998;83:3071.

39. Pinna G, Meinhold H, Hiedra L, et al. Elevated 3,5-diiodothyronine concentrations in the sera of patients with nonthyroidal illness and brain tumors. J Clin Endocrinol Metab 1997;82: 1535.

40. Chopra IJ, Wu SY, Chua Teco GN, et al. A radioimmunoassay of 3,5,3′-triiodothyronine sulfate: studies in thyroidal and nonthyroidal diseases, pregnancy, and neonatal life. J Clin Endocrinol Metab 1992;75:189.

41. Santini F, Chiovato L, Bartalena L, et al. Study of serum 3,5,3′-triiodothyronine sulfate concentration in patients with systemic nonthyroidal illness. Eur J Endocrinol 1996;134:45.

42. Chopra IJ, Santini F, Hurd RE, et al. A radioimmunoassay for measurement of thyroxine sulfate. J Clin Endocrinol Metab 1993;76:145.

43. Meinhold H, Gramm HJ, Meissner W, et al. Elevated serum diiodotyrosine (DIT) in severe infections and sepsis: DIT, a possible new marker of leukocyte activity. J Clin Endocrinol Metab 1991;72:945.

44. Chamba A, Neuberger J, Strain A, et al. Expression and function of thyroid hormone receptor variants in normal and chronically diseased human liver. J Clin Endocrinol Metab 1996;81:360.

45. Beigneux AP, Moser AH, Shigenaga JK, et al. Sick euthyroid syndrome is associated with decreased TR expression and DNA binding in mouse liver. Am J Physiol Endocrinol Metab 2003; 284:E228.

46. Osburn RC, Mayers EA, Rodbard D, et al. Adaptation to hypocaloric feeding: physiological significance of the fall in serum T3 as measured by pulse wave arrival time (QKd). Metabolism 1983;32:9.

47. Croxson MS, Ibbertson HK. Low serum triiodothyronine (T3) and hypothyroidism in anorexia nervosa. J Clin Endocrinol Metab 1977;44:167.

48. Brent GA, Hershman JM, Reed AW, et al. Serum angiotensin-converting enzyme in severe nonthyroidal illness associated with low serum thyroxine concentration. Ann Intern Med 1984;100:680.

49. Seppel T, Becker A, Lippert F, et al. Serum sex hormone-binding globulin and osteocalcin in systemic nonthyroidal illness associated with low thyroid hormone concentrations. J Clin Endocrinol Metab 1996;81:1663.

50. Dasmahapatra A, Cohen MP, Grossman SD, et al. Erythrocyte sodium/potassium adenosine triphosphatase in thyroid disease and nonthyroidal illness. J Clin Endocrinol Metab 1985;61:110.

51. Brandt MR, Kehlet H, Skovsted L, et al. Rapid decrease in plasma triiodothyronine during surgery and epidural analgesia independent of afferent neurogenic stimuli and cortisol. Lancet 1976;2:1333.

52. Jakobs TC, Mentrup B, Schnutzler C, et al. Proinflammatory cytokines inhibit the expression and function of human type I 5′-deiodinase in HepG2 hepatocarcinoma cells. Eur J Endocrinol 2002;146:559.

53. Yu J, Koenig RJ. Regulation of hepatocyte thyroxine 5′-deiodinase by T3 and nuclear receptor coactivators as a model of the sick euthyroid syndrome. J Biol Chem 2000;275:38296.

54. Nagaya T, Fujieda M, Otsuka G, et al. A potential role of activated NF-kB in the pathogenesis of the sick euthyroid syndrome. J Clin Invest 2000;106:393.

55. Lim C-F, Bernard BF, de Jong M, et al. A furan fatty acid and indoxyl sulfate are the putative inhibitors of thyroxine hepatocyte transport in uremia. J Clin Endocrinol Metab 1993;76: 318.

56. Everts ME, Lim CF, Moerings EPCM, et al. Effects of a furan fatty acid and indoxyl sulfate on thyroid hormone uptake in cultured anterior pituitary cells. Am J Physiol 1995; 268:E974.

57. Lim C-F, Docter R, Visser TJ, et al. Inhibition of thyroxine transport into cultured rat hepatocytes by serum of nonuremic critically ill patients: effects of bilirubin and nonesterified fatty acids. J Clin Endocrinol Metab 1993;76:1165.

58. Vos RA, Jong M de, Bernard BF, et al. Impaired thyroxine and 3,5,3′-triiodothyronine handling by rat hepatocytes in the presence of serum of patients with nonthyroidal illness. J Clin Endocrinol Metab 1995;80:2364.

59. Hennemann G, Everts ME, de Jong M, et al. The significance of plasma membrane transport in the bioavailability of thyroid hormone. Clin Endocrinol (Oxf) 1998;48:1.

60. Lim C-F, Docter R, Krenning EP, et al. Transport of thyroxine into cultured hepatocytes: effects of mild non-thyroidal illness and caloric restriction in obese subjects. Clin Endocrinol (Oxf) 1994;40:79.

61. Van der Heijden JTM, Docter R, Van Toor H, et al. Effect of caloric deprivation on the thyroid hormone tissue uptake and the generation of the low-T3syndrome. Am J Physiol 1986; 251:E156.

62. Hennemann G, Docter R, Friesema ECH, et al. Plasma membrane transport of thyroid hormones and its role in thyroid hormone metabolism and bioavailability. Endocr Rev 2001; 22:451.

63. De Jong M, Docter R, Bernard HF, et al. T4 uptake into the perfused rat liver and liver T4 uptake in humans are inhibited by fructose. Am J Physiol 1994;266:E768.

64. Jirasakuldech B, Schussler GC, Yap MG, et al. A characteristic serpin cleavage product of thyroxine-binding globulin appears in sepsis sera. J Clin Endocrinol Metab 2000;85:3996.

65. Afandi B, Schussler GC, Arafeh AH, et al. Selective consumption of thyroxine-binding globulin during cardiac bypass surgery. Metabolism 2000;49:270.

66. Mendel CM, Laughton CW, McMahon FA, et al. Inability to detect an inhibitor of thyroxine-serum protein binding in sera from patients with nonthyroidal illness. Metabolism 1991;40:491.

67. Chopra IJ, Huang T-S, Beredo A, et al. Serum thyroid hormone binding inhibitor in nonthyroidal illnesses. Metabolism 1986;35:152.

68. Willcox RB, Nelson JC, Tomei RT. Heterogeneity in affinities of serum proteins for thyroxine among patients with nonthyroidal illness as indicated by the serum free thyroxine response to serum dilution. Eur J Endocrinol 1994;131:9.

69. Mendel CM, Frost PH, Cavalieri RR. Effect of free fatty acids on the concentration of free thyroxine in human serum: the role of albumin. J Clin Endocrinol Metab 1986;63:1394.

70. Mendel CM, Frost PH, Kunitaka ST, et al. Mechanism of the heparin-induced increase in the concentration of free thyroxine in plasma. J Clin Endocrinol Metab 1987;65:1259.

71. Wolf M, Hansen N, Greten H. Interleukin-1β, tumor necrosis factor-α, and interleukin 6 decrease nuclear thyroid hormone receptor capacity in a liver cell line. Eur J Endocrinol 1994; 131:307.

72. Moordian AD, Reed RL, Osterweil D, et al. Decreased serum triiodothyronine is associated with increased concentrations of tumor necrosis factor. J Clin Endocrinol Metab 1990;71:1239.

73. Chopra IJ, Sakane S, Chua Teco GN. A study of the serum concentration of tumor necrosis factor-α in thyroidal and nonthyroidal illness. J Clin Endocrinol Metab 1991;72:1113.

74. Hashimoto H, Igarashi N, Yachie A, et al. The relationship between serum levels of interleukin-6 and thyroid hormone in children with acute respiratory infection. J Clin Endocrinol Metab 1994;78:288.

75. Boelen A, Platvoet-ter Schiphorst MC, Wiersinga WM. Relationship between serum 3,5,3′-triiodothyronine and serum interleukin-8, interleukin-10 or interferon-γ in patients with nonthyroidal illness. J Endocrinol Invest 1996;19:480.

76. Boelen A, Platvoet-ter Schiphorst M, Wiersinga WM. Association between serum interleukin-6 and serum 3,5,3′-triiodothyronine in nonthyroidal illness. J Clin Endocrinol Metab 1993; 77:1695.

77. Bartalena L, Brogioni S, Grasso L, et al. Relationship of the increased serum interleukin-6 concentration to changes of thyroid function in nonthyroidal illness. J Endocrinol Invest 1994; 17:269.

78. Davies PH, Black EG, Sheppard MC, et al. Relation between serum interleukin-6 and thyroid hormone concentrations in 270 hospital in-patients with non-thyroidal illness. Clin Endocrinol (Oxf) 1996;44:199.

79. Boelen A, Platvoet-ter Schiphorst MC, Wiersinga WM. Soluble cytokine receptors and the low 3,5,3′-triiodothyronine syndrome in patients with nonthyroidal disease. J Clin Endocrinol Metab 1995;80:971.

80. Poll T van der, Romijn JA, Wiersinga WM, et al. Tumor necrosis factor: a putative mediator of the sick euthyroid syndrome in man. J Clin Endocrinol Metab 1990;71:1567.

81. Corssmit EPM, Heijligenberg R, Endert E, et al. Acute effects of interferon-α administration on thyroid hormone metabolism in healthy men. J Clin Endocrinol Metab 1995;80: 3140.

82. Stouthard JML, van der Poll T, Endert E, et al. Effects of acute and chronic interleukin-6 administration on thyroid hormone metabolism in humans. J Clin Endocrinol Metab 1994;79: 1342.

83. Metz J de, Romijn JA, Endert E, et al. Administration of interferon-γ in healthy subjects does not modulate thyroid hormone metabolism. Thyroid 2000;10:87.

84. Van der Poll T, van Zee KJ, Endert E, et al. Interleukin-1 receptor blockade does not affect endotoxin-induced changes in plasma thyroid hormone and thyrtropin concentrations in man. J Clin Endocrinol Metab 1995;80:1341.

85. Van der Poll T, Endert E, Coyle SM, et al. Neutralization of TNF does not influence endotoxin-induced changes in thyroid hormone metabolism in humans. Am J Physiol 1999; 276:R357.

86. Boelen A, Platvoet-ter Schiphorst MC, Wiersinga WM. Immunoneutralization of interleukin-1, tumor necrosis factor, interleukin-6 or interferon does not prevent the LPS-induced sick euthyroid syndrome in mice. J Endocrinol 1997;153: 115.

87. Boelen A, Maas MAW, Lowik CWGM, et al. Induced illness in interleukin-6 (IL-6) knock-out mice: a causal role of IL-6 in the development of the low 3,5,3′-triiodothyronine syndrome. Endocrinology 1996;137:5250.

88. Maturlo SJ, Rosenbaum RL, Pan C, et al. Variable thyrotropin response to thyrotropin-releasing hormone after small decreases in plasma free thyroid hormone concentrations in patients with nonthyroidal diseases. J Clin Invest 1980;66:451.

89. Diano S, Naftolin F, Goglin F, et al. Fasting-induced increase in type II iodothyronine deiodinase activity and messenger ribonucleic acid level is not reversed by thyroxine in the rat hypothalamus. Endocrinology 1998;139:2879.

90. Fekete C, Gereben B, Doleschall M, et al. Lipopolysaccharide induces type 2 iodothyronine deiodinase in the mediobasal hypothalamus: implications for the nonthyroidal illness syndrome. Endocrinology 2004;145:1649.

91. Kakucska I, Qi Y, Lechan RM. Changes in adrenal status affect hypothalamic thyrotropin-releasing hormone gene expression in parallel with corticotropin-releasing hormone. Endocrinology 1995;136:2795.

92. Kondo K, Harbuz MS, Levy A, et al. Inhibition of the hypothalamic-pituitary-thyroid axis in response to lipopolysaccharide is independent of changes in circulating corticosteroids. Suppression of thyrotropin-releasing hormone gene expression by interleukin-1 beta in the rat: implications for nonthyroidal illness. Neuroimmunomodulation 1997;4:188.

93. Mihaly E, Fekete C, Tatro JB, et al. Hypophysiotropic thyrotropin releasing hormone-synthesizing neurons in the human hypothalamus are innervated by neuropeptide Y, agouti-related protein, and alpha-melanocyte-stimulating hormone. J Clin Endocrinol Metab 2000;85:2596.

94. Fekete C, Sarkar S, Rand WM, et al. Agouti-related protein (AGRP) has a central inhibitory action on the hypothalamic-pituitary thyroid (HPT) axis; comparisons between the effect of AGRP and neuropeptide Y on energy homeostasis and the HPT axis. Endocrinology 2002;143:3846.

95. Fekete C, Legradi G, Mihaly E, et al. Alpha-melanocyte-stimulating hormone is contained in nerve terminals innervating thyrotropin-releasing hormone-synthesizing neurons in hypothalamic paraventricular nucleus and prevents fasting-induced suppresion of prothyrotropin-releasing hormone gene expression. J Neurosci 2000;20:1550.

96. Legradi G, Emerson CH, Ahima RS, et al. Leptin prevents fasting-induced suppression of prothyrotropin-releasing hormone messenger ribonucleic acid in neurons of the hypothalamic paraventricular nucleus. Endocrinology 1997;138:2569.

97. Legradi G, Emerson CH, Ahima RS, et al. Arcuate nucleus ablation prevents fasting-induced suppression of proTRH mRNA in the hypothalamic paraventricular nucleus. Neuroendocrinology 1998;68:89.

98. Chan JL, Heist K, DePaoli AM, et al. The role of falling leptin levels in the neuroendocrine and metabolic adaptation to short-term starvation in healthy men. J Clin Invest 2003;111: 1409.

99. Bornstein SR, Torpy DJ, Chrousos GP, et al. Leptin levels are elevated despite low thyroid hormone levels in the “euthyroid sick” syndrome. J Clin Endocrinol Metab 1997;82: 4278.

100. Van den Berghe G, Wouters P, Carlsson L, et al. Leptin levels in protracted critical illness: effects of growth hormone-secretagogues and thyrotropin-releasing hormone. J Clin Endocrinol Metab 1998;83:3062.

101. Fliers E, Unmehopa U, Manniesing S, et al. Decreased neuropeptide Y (NPY) expression in the infundibular nucleus of patients with nonthyroidal illness. Peptides 2001;22:459.

102. Kakucska I, Romero LI, Clark BD, et al. Suppression of thyrotropin-releasing hormone gene expression by interleukin-1b in the rat: implications for nonthyroidal illness. Neuroendocrinology 1994;59:127.

103. Wassen FNJS, Moerings EPCM, van Toor H, et al. Effects of interleukin-1b on thyrotropin secretion and thyroid hormone uptake in clutured rat anterior pituitary cells. Endocrinology 1996;137;1591.

104. Arem R, Thornby JI, Deppe SA. Comparison of thyroid hormone and cortisol measurements with APACHE II and TISS scoring systems as predictors of mortality in the medical intensive care unit. J Intens Care Med 1997;12:12.

105. Spencer CA. Clinical utility and cost-effectiveness of sensitive thyrotropin assays in ambulatory and hospitalized patients. Mayo Clin Proc 1988;63:1214.

106. Franklyn JA, Black EG, Betteridge J, et al. Comparison of second and third generation methods for measurement of serum thyrotropin in patients with overt hyperthyroidism, patients receiving thyroxine therapy, and those with nonthyroidal illness. J Clin Endocrinol Metab 1994;78:1368.

107. Burmeister LA. Reverse T3 does not reliably differentiate hypothyroid sick syndrome from euthyroid sick syndrome. Thyroid 1995;5:435.

108. Hooper MJ. Diminished TSH secretion during acute non-thyroidal illness in untreated primary hypothyroidism. Lancet 1976;1:48.

109. Nair KS, Halliday D, Ford GC, et al. Effect of triiodothyronine on leucine kinetics, metabolic rate, glucose concentration and insulin secretion rate during two weeks of fasting in obese women. Int J Obesity 1989;13:487.

110. Gardner DF, Kaplan MM, Stanley CA, et al. Effect of triiodothyronine replacement on the metabolic and pituitary responses to starvation. N Engl J Med 1979;300:579.

111. Burman KD, Wartofsky L, Dinterman RE, et al. The effect of T3 and reverse T3 administration on muscle protein catabolism during fasting as measured by 3-methylhistidine excretion. Metabolism 1979;28:805.

112. Byerley LO, Heber D. Metabolic effects of triiodothyronine replacement during fasting in obese subjects. J Clin Endocrinol Metab 1996;81:968.

113. Brent GA, Hershman JM. Thyroxine therapy in patients with severe nonthyroidal illnesses and low serum thyroxine concentration. J Clin Endocrinol Metab 1986;63:1.

114. Becker RA, Vaughan GM, Ziegler MG, et al. Hypermetabolic low triiodothyronine syndrome of burn injury. Crit Care Med 1982;10:870.

115. Acker CG, Singh AR, Flick RP, et al. A trial of thyroxine in acute renal failure. Kidney Int 2000;57:293.

116. Lim VS, Tsalikian E, Flanigan MJ. Augmentation of protein degradation by L-triiodothyronine in uremia. Metabolism 1989;38:1210.

117. Little JS. Effect of thyroid hormone supplementation on survival after bacterial infection. Endocrinology 1985;117: 1431.

118. Chopra IJ, Huang TS, Boado R, et al. Evidence against benefit from replacement doses of thyroid hormone in nonthyroidal illness (NTI): studies using turpentine oil-injected rats. J Endocrinol Invest 1987;10:559.

119. Dulchavaky SA, Kennedy PR, Geller ER, et al. T3 preserves respiratory function in sepsis. J Trauma 1991;31:753.

120. Klemperer JD, Klein I, Gomez M, et al. Thyroid hormone treatment after coronary-artery bypass surgery. N Engl J Med 1995;333:1522.

121. Klemperer JD, Klein IL, Ojaman K, et al. Triiodothyronine therapy lowers the incidence of atrial fibrillation after cardiac operations. Ann Thorac Surg 1996;61:1323.

122. Bennett-Guerrero E, Jimenez JL, White WD, et al. Cardiovascular effects of intravenous triiodothyronine in patients undergoing coronary artery bypass graft surgery: a randomized, double-blind, placebo-controlled trial. JAMA 1996;275:687.

123. Mullis-Jansson SL, Argeneiano M, Corwin S, et al. A randomized double-blind study of the effect of triiodothyronine on cardiac function and morbidity after coronary-bypass surgery. J Thorac Cardiovasc Surg 1999;117:1128.

124. Chowdhury D, Parnell V, Ojamaa K, et al. Usefulness of triiodothyronine (T3) treatment after surgery for complex congenital heart disease in infants and children. Am J Cardiol 1999;84:1107.

125. Bettendorf M, Schmidt KG, Grulich-Henn J, et al. Triiodothyronine treatment in children after cardiac surgery: a double-blind, randomised, placebo-controlled study. Lancet 2000; 356:529.

126. Van den Berghe G, Wouters P, Weekers F, et al. Reactivation of pituitary hormone release and metabolic improvement by infusion of growth hormone-releasing peptide and thyrotropin-releasing hormone in patients with protracted critical illness. J Clin Endocrinol Metab 1999;84:1311.

127. Iitaka M, Kawasaki S, Sakurai S-I, et al. Serum substances that interfere with thyroid hormone assays in chronic renal failure. Clin Endocrinol (Oxf) 1998;48:739.

128. Reinhardt W, Misch C, Jockenhövel F, et al. Triiodothyronine (T3) reflects renal graft function after renal transplantation. Clin Endocrinol (Oxf) 1997;46:563.

129. Kaptein EM. Thyroid hormone metabolism and thyroid diseases in chronic renal failure. Endocr Rev 1996;17:45.

130. LoPresti JS, Fried JC, Spencer CA, et al. Unique alterations of thyroid hormone indices in the acquired immunodeficiency syndrome (AIDS). Ann Intern Med 1989;110:970.

131. Hommes MJT, Romijn JA, Endert E, et al. Hypothyroid-like regulation of the pituitary–thyroid axis in stable human immunodeficiency virus infection. Metabolism 1993;42:556.

132. Hirschfeld S, Lane L, Cutler GB, et al. Thyroid abnormalities in children infected with human immunodeficiency virus. J Pediatr 1996;128:70.

133. Grappin M, Piroth L, Verges B, et al. Increased prevalence of subclinical hypothyroidism in HIV patients treated with highly active antiretroviral therapy. AIDS 2000;14:1070.

134. Calza L, Manfredi R, Chiodo F. Subclinical hypothyroidism in HIV-infected patients receiving highly active antiretroviral therapy. J Acquired Immune Deficiency Syndromes 2002;31: 361.

135. Beltran S, Lescure FX, Desailloud R, et al. Increased prevalence of hypothyroidism among human immunodefiency virus-infected patients: a need for screening. Clin Inf Dis 2003; 37:579.

136. Renard E, Fabre J, Paris F, et al. Syndrome of body fat redistribution in HIV-1-infected patients: relationships to cortisol and catecholamines. Clin Endocrinol (Oxf) 1999;51:223.

137. Koutkia P, Mylonakis E, Levin RM. Human immunodeficiency virus infection and the thyroid. Thyroid 2002;12: 577.

138. Mason J, Southwick S, Yehuda R, et al. Elevation of serum free triiodothyronine, total triiodothyronine, thyroxine-binding globulin, and total thyroxine levels in combat-related posttraumatic stress disorder. Arch Gen Psychiatry 1994;51: 629.

139. Nader S, Warner MD, Doyle S, et al. Euthyroid sick syndrome in psychiatric inpatients. Biol Psychiatry 1996;40:1288.

140. Arem R, Cusi K. Thyroid function testing in psychiatric illness. Usefulness and limitations. Trends Endocrinol Metab 1997; 8:282.

141. Kirkegaard C, Faber J. The role of thyroid hormones in depression. Eur J Endocrinol 1998;138:1.

142. Alkemade A, Unmehopa UA, Brouwer JP, et al. Decreased thyrotropin-releasing hormone gene expression in the hypothalamic paraventricular nucleus of patients with major depression. Mol Psychiatry 2003;8:838.

143. Aronson R, Offman H, Joffe RT, et al. Triiodothyronine augmentation in the treatment of refractory depression. A meta-analysis. Arch Gen Psychiatry 1996;53:842.

144. Altshuler LL, Bauer M, Frye MA, et al. Does thyroid supplementation accelerate tricyclic antidepressant response? A review and meta-analysis of the literature. Am J Psych 2001;158: 1617.

145. Fliers E, Appelhof BC, Brouwer JP, et al. Effect of triiodothyronine (T3) addition to paroxetine in major depressive disorder: a randomized clinical trial. In: Program and abstracts of the Endocrine Society's 85thAnnual Meeting; 2003; Philadelphia, PA. Abstract S-19.2.