Euthyroid Sick Syndrome: Is It a Misnomer?

THE TERM euthyroid sick syndrome (ESS) identifies abnormalities in thyroid function tests observed in patients with systemic nonthyroidal illnesses (NTIs) and those undergoing surgery or fasting (1, 2). The term nonthyroidal illness syndrome (NTIS) has also been employed to describe these abnormalities (3). These abnormalities result from variable, usually reversible, disturbances in the hypothalamo-pituitary-thyroid axis, thyroid hormone binding to serum proteins, tissue uptake of thyroid hormones, and/or thyroid hormone metabolism. Several recent reviews have addressed these issues (3–6). I shall focus mainly on the clinical diagnosis, significance, and treatment of ESS. Abnormalities of thyroid function in NTIS have been classified as 1) low T3 syndrome, 2) low T3-low T4 syndrome, 3) high T4 syndrome, and 4) other abnormalities (7). Although serum concentrations of total T3 and T4 are now measured routinely by similar RIAs, several methods have been employed for measurement of the small, biologically active, free fraction of T3 and T4 (8–14). Most workers in the field view the measurement of free thyroid hormones by equilibrium dialysis as the gold standard, and the ultrafiltration method is comparable or a close second. Until recently, tracer equilibrium dialysis was believed to be the most accurate procedure for measurement of free T3 or T4. This is expected to be increasingly replaced by the newer, more accurate, equilibrium dialysis/RIA (12, 14); reasonably priced kits are available commercially for free T4 measurement by this procedure (Nichols Diagnostics, San Juan Capistrano, CA) and should be available soon for free T3 measurement. A detailed discussion of methods of measurement of free thyroid hormones is beyond the scope of this minireview, and the reader is referred to several studies comparing available procedures (9, 10, 13). The analog methods for free thyroid hormone measurements are popular in several countries outside of the United States. These methods yield free thyroid hormone readings in NTI that are similar to those from the index method and different from those by the tracer equilibrium dialysis or equilibrium dialysis/RIA procedures (9, 10, 13). For this review, I have relied mainly on free thyroid hormone levels measured by the newer equilibrium dialysis/RIA procedure when data are available, or those measured by tracer dialysis and/or ultrafiltration procedures. Low serum total T3 is the most common abnormality in NTI. It is observed in about 70% of hospitalized patients. Serum total T3 may vary from undetectable to normal in patients with systemic illness, and the mean value is approximately 40% of the normal level. The serum free T3 concentration as measured by direct equilibrium dialysis/RIA [free T3(D)] is also decreased, but less severely, and the mean value is approximately 60% that of normal (14). The serum free T3 concentration measured by ultrafiltration is either normal or reduced (11). Low values were observed in patients given dopamine. It is curious that free T3 was often low in NTI when measured by equilibrium dialysis (14) and typically normal when measured by ultrafiltration (11). The discrepancy is possibly explained by the lower total T3 and more severe NTI in patients determined by equilibrium dialysis than by ultrafiltration. Free T3 in systemic illness has also been measured by a variety of other techniques and has been found to be low, but the accuracy of these procedures is questionable in NTI patients (8, 10, 15). When measured, the daily production rate (PR) of T3 is decreased in NTI (16, 17), which supports the finding in NTI of low free T3(D). Serum total T4 and free T4(D) and, when measured, daily PR-T4 are normal in the low T3 syndrome (12, 16, 18, 19). A decreased serum free T3(D) concentration and PR-T3 at a time when the serum free T4(D) concentration and PR-T4 are normal reflect decreased conversion of T4 to T3 in NTI (12, 16, 18, 19). The serum concentration of rT3 is increased in NTI, except in renal failure (20). However, daily PR-rT3 is normal, and the increase in the serum rT3 level is related mainly to the delayed MCR of rT3, which is predominantly due to decreased activity of the type I iodothyronine 5′-monodeiodinase (5′-MDI) in tissues (16); 5′-MDI deiodinates T4 to T3 and rT3 to 3,3′-diiodothyronine (T2) (21). The low T3 and low T4 syndrome is observed in severely ill, frequently moribund, patients, usually admitted to medical intensive care units. Low serum total T4 correlates with a bad prognosis (22). The serum concentration of free T4, as measured by equilibrium dialysis/RIA [free T4(D)], is normal in most NTI patients with low total T4 (12). Interestingly, total T4 is often low in NTI patients even when their serum concentration of immunoassayable T4-binding globulin (TBG) is clearly normal (8). However, the free T4 index is frequently low in these patients (7, 9, 10). This combination of findings in NTI has been explained by the presence in the circulation of an inhibitor of serum (and resin) binding of thyroid hormones (15, 23). The nature of the inhibitor is not known. We and others have considered a role for nonesterified fatty acids in some cases, especially when serum albumin is low (15, 24, 25). An inhibitor of serum binding of thyroid hormones is present in tissues (26), and its nature or its leakage into the circulation are not known. Some investigators do not agree with the existence of such an inhibitor in the sera of NTI patients (27). On the basis of experiments involving mixing of normal sera with NTI sera, Mendel et al. (27) were unable to document the existence of an inhibitor in NTI. They suggested that diminished serum binding of T4 in NTI patients with normal immunoassayable TBG is a result of high level of desialylated TBG (27). This may indeed be the case, but direct measurements of desialylated TBG were not performed. Interestingly, however, the researchers observed that desialylated TBG has markedly decreased avidity for T4, but its avidity for T3 remains unchanged (27). Therefore, the finding of clearly decreased serum binding of T3, evidenced by a markedly elevated dialyzable fraction of T3, in several NTI patients with normal immunoassayable TBG (8) supports the possibility of a thyroid hormone binding inhibitor in NTI. When present, a decreased serum concentration and/or affinity of thyroid hormone binding proteins, especially TBG, can explain the findings of low total T4 and normal free T4(D) in NTI. Decreased serum binding of T4 in NTI is associated with an increased MCR, which, too, contributes to a decreased serum concentration of total T4. Interestingly, however, the increase in the MCR of T4 in NTI is not as much as expected from the degree of reduction in serum binding of thyroid hormones (4). The serum free T4 concentration is low in NTI patients treated with dopamine and corticosteroids, which decrease serum TSH levels (11, 28–30). Besides low TSH, factors that may contribute to the low T4 of NTI include abnormalities in TSH secretion, decreased biological activity of TSH, and diminished thyroidal response to TSH (31, 32). High serum total T4 is seen in some NTI patients, who have elevated serum concentrations of TBG. Serum TBG is elevated in acute intermittent porphyria (33) and several liver diseases, including chronic hepatitis and primary biliary cirrhosis (34). The serum concentration of free T4(D) is normal in these patients in the absence of thyroid disease. Serum total T3 may be normal, but free T3(D) or the free T3 index is low normal or low as in other patients with NTI. The serum concentration of rT3 is elevated in NTI patients with high T4. Both serum total and free T4(D) concentrations are often increased in NTI patients treated with amiodarone and iodinated radiocontrast agents, e.g. iopanoic acid and ipodate used for oral cholecystography (13, 35, 36). These agents decrease hepatic uptake of T4 and 5′-monodeiodination of T4 (to T3) and, in addition, may precipitate hyperthyroidism in patients with autonomous thyroid nodules by invoking the Jod Basedow phenomenon (37, 38). The effect of a single dose of oral cholecystography agents on serum T4 typically lasts less than 24 h (36, 37, 39). NTI patients with high total and free T4(D), especially those who have ingested iodine-containing agents, should be followed carefully for the appearance of typical hyperthyroidism (38). Serum T3 may be normal or even low initially because of the effects of the drug and/or NTI on peripheral conversion of T4 to T3, and it may increase dramatically during follow-up. The serum concentration of free T4(D) is elevated in NTI patients given heparin (40). This is an in vitro artifact explained by displacement of T4 from binding proteins by fatty acids generated from the action of lipase(s) on plasma triglycerides. Total T4 and the free T4 index are normal in these patients, who are clinically euthyroid. Infection with human immunodeficiency virus (HIV) produces unusual alterations in thyroid function, including increases in T4 and TBG, paradoxical decreases in rT3 and the rT3/T4 ratio, and the maintenance of a normal T3 and T3/T4 ratio even in severely ill patients. Serum T3 decreases, however, in critically ill patients with HIV and pneumocystis infection (41). The basis for the differences in thyroid hormone abnormalities in HIV compared to those in other NTIs is not known. Some factors that may contribute to major abnormalities of the NTIS are listed in Fig. 1. They have recently been critically reviewed (3–6). There is evidence for decreased conversion of T4 to T3 in extrathyroidal tissues in the NTIS (16–19), and this may, in turn, be related to decreased activity and/or concentration of 5′-MDI (42). 5′-MDI has now been cloned in the rat and man, and it turns out that it belongs to a small group of selenocysteine-containing proteins (43, 44). No data are available on the tissue content of the 5′-MDI protein or its messenger ribonucleic acid in human NTIS. However, the hepatic content of 5′-MDI protein was decreased in the fasting rat, studied as a model of NTIS (42). Diminution in the uptake of T4 by tissues can also explain the decreased generation of T3 in tissues (5). However, this abnormality should be associated with an elevated serum concentration of free T4(D), which is clearly not the case (see above) (45). One could argue that decreased T4 to T3 conversion in NTI should also be associated with increased free T4(D). However, T4 is metabolized not just by 5′-MDI, but also by type III iodothyronine deiodinase, conjugation and side-chain alteration (21, 46); these alternate routes of T4 metabolism are not known to be impaired in NTI and may compensate for T4 not metabolized by 5′-MDI. Some factors that may contribute to major abnormalities of NTIS. The reader is referred to previous reviews (3–6) for detailed discussion of these factors. Alterations in serum binding of thyroid hormones is clearly an important factor contributing to changes in thyroid hormone levels in NTI (see above). Serum albumin binds compounds, e.g. fatty acids, that are capable of displacing thyroid hormones from TBG. The fall in serum albumin in NTI enhances the activity of such low affinity competitors of T4 on TBG (15, 24, 25). Additionally, much has been written on abnormalities in the synthesis, secretion, structure, regulation, and effectiveness of TSH in NTI (3–6, 31, 32). There has been much interest recently in the roles of cytokines in the pathogenesis of the NTI (3). However, their significance remains unclear. Proinflammatory cytokines [tumor necrosis factor-α (TNFα), interleukins (e.g. IL-1 and IL-6), and interferon-γ], when administered to man or experimental animals, have caused changes in thyroid function tests that resemble NTIS (47–50). However, humans or animals so treated manifest substantial systemic illness, and it is unclear whether the thyroid hormone changes observed are due to the sickness induced by cytokines or the cytokines per se. In this respect, it is curious that lipopolysaccharide-induced NTIS in mice, although associated with increases in circulating TNFα and IL-6, was not prevented by immunoneutralization of IL-1 receptor, TNFα, IL-6, or interferon-γ (51) (Wiersinga, W. M., personal communication). Abnormal thyroid function tests are observed at least as frequently in systemic NTIS as in thyroid diseases (4, 52, 53). Thyroid function abnormalities in NTI may at times mimic and at other times mask the biochemical abnormalities observed in true thyroid disease. Furthermore, the severity and the nature of changes in thyroid function test have implications for the prognosis of the systemic illness. Thus, a low serum T3 level predicts increased mortality form liver cirrhosis, advanced congestive heart failure (54, 55), and possibly several other systemic illnesses. Similarly, low total T4 is associated with increased mortality from systemic illness, and those patients with low T4 who have very low serum T3 levels have the worst prognosis (22, 55, 56). Previous studies have suggested that there exist tissue factors that can inhibit the binding of thyroid hormones to serum proteins and the ability of polymorphonuclear leukocytes to phagocytose Escherichia coli (26, 57). However, it is not known whether the two effects are due to the same or even similar factors and whether the thyroid hormone binding inhibitor considered above in serum is similar to that extracted from tissues. In any case, leakage of tissue elements in the circulation in systemic illnesses may explain the correlation between the fall in serum T4 and the increased mortality in NTI; this requires further study. It can be challenging to establish the diagnosis of thyroid disease in patients manifesting NTIS. The difficulty exists in hyperthyroid patients who may exhibit normal total T4 and T3 on account of a diminution in serum binding of thyroid hormones. However, serum free T4 and free T3 determinations by equilibrium dialysis/RIA (12) or ultrafiltration (45) should yield appropriately diagnostic elevated values. The free T4 index and analog methods for free T4 determination frequently yield low values in NTI and should be interpreted with caution. Serum TSH measured by the ultrasensitive RIA is typically undetectable (<0.10 mU/mL) in hyperthyroidism, whereas it is undetectable in less than 7% patients with NTI, usually those who have been treated with dopamine or corticosteroids (30). Definite diagnosis of hypothyroidism can also be difficult in the setting of NTI. However, primary hypothyroidism is a strong possibility if serum TSH is above 25–30 μU/mL. Serum TSH is supranormal in ∼12% patients with NTI, and it is above 20 μU/mL in less than 3% of patients (30, 58). A subnormal free T4(D) concentration in the absence of treatment of NTI patients with TSH-suppressive drugs, e.g. dopamine, corticosteroids, and anticonvulsants (e.g. phenytoin and carbamozeprine) is strongly suggestive of hypothyroidism. As Hashimoto’s thyroiditis is a common cause of hypothroidism in our patient population, findings of goiter and positive antithyroid antibodies (e.g. thyroid peroxidase and thyoglobulin autoantibodies) in serum are strong points favoring the diagnosis of primary hypothyroidism. An elevated serum rT3 level argues against the diagnosis of hypothyroidism, when serum TSH is more than 10 μU/mL (59). As an individual test, serum rT3 does not help in the diagnosis of hypothyroidism in NTI patients. Low, normal, and high serum rT3 values are observed in patients with TSH values varying between low and 10 μU/mL (normal, 0.5–5.0μ U/mL). However, likely hypothyroid patients with serum TSH levels above 20 μU/mL do not demonstrate supranormal rT3 (59). It is prudent not to rely solely on any one thyroid function test in the setting of NTI, and a combination of tests should be considered in separating primary hypothyroid from euthyroid patients with the NTIS. The diagnosis of secondary (or tertiary) hypothyroidism may require additional work-up. The serum TSH level may be low, normal, or minimally elevated in secondary hypothyroidism. Plasma cortisol is clearly elevated or high normal, whereas PRL and gonadotropin levels are typically normal in NTI without a specific pituitary or hypothalamic disease. On the other hand, decreased plasma cortisol and gonadotropin levels and elevated PRL levels support a diagnosis of a central (pituitary or hypothalamic) lesion as the basis for secondary hypothyroidism (60, 61). In view of several above-mentioned complexities, it is often reasonable to wait for a week or so after recovery from an acute NTI before evaluating thyroid status. Many believe that patients with NTIS are metabolically euthyroid even though serum levels of the most active thyroid hormone, T3, are clearly in the hypothyroid range. A normal serum TSH in most NTI patients is clearly an important element in this belief. However, several studies demonstrating abnormalities in the synthesis, secretion, glycosylation, regulation, and/or effectiveness of TSH in NTI (see above) lead one to question the normalcy of TSH in the face of low serum levels of active thyroid hormone (T3) in NTI. The transient increase in serum TSH during recovery from NTI suggests that TSH is suppressed in an illness (62). Pituitary TSH suppression may be related to the stress of an illness, and the resulting elevated cortisol and catecholamine levels and associated caloric deprivation (4, 62, 63). The basis for the apparent euthyroid status in NTI remains unclear, but one or more of the following explanations may be considered: 1) a moderate degree and a short duration of reduction in thyroid hormone (T3) levels, 2) insensitivity of the clinical diagnosis of (mild) hypothyroidism, and 3) increased sensitivity of body tissues to T3. Although available data are limited and conflicting, a number of studies support the opposite viewpoint. Thus, oxygen consumption in response to T3 has been found to be decreased (not increased) in fasting rats compared to fed rats (64). There is also information suggesting a decreased effect of low T3 on TSH in fasting subjects with the low T3 syndrome (65, 66). Similarly, there is a decrease (not an increase) in the binding of T3 to nuclear receptors to T3 in diabetes mellitus and fasting in the rat (67, 68). There is also evidence for a diminution in thyroid hormone effects at the postreceptor level (69). One study, however, has suggested an increase in T3 receptor number and affinity in NTIS (70). Studies in the 1970s suggested low T3 to be protective against protein breakdown in fasting (71, 72). However, a recent study was unable to document hypercatabolic effects of T3 in fasting obese subjects (73). Finally, 4) there may exist in the sera of NTI patients high levels of thyromimetic compounds other than T3. A high serum concentration of T3 sulfate in NTI is of interest in this regard (74, 75). Increased exposure of body tissues to 3,5,3′-triiodothyroacetic acid has also been suggested in NTI (76). Overall, it seems that although several patients with NTI may be euthyroid because of a short duration of NTIS, normal free T3, and/or a of there are especially those with a NTI (and those who manifest low free T4 by who may indeed be hypothyroid and may from treatment with thyroid This is by data that tissues of patients from NTI much decreased levels of thyroid hormones compared to tissues of subjects who and that the degree of thyroid hormone from one to The of secondary hypothyroidism in NTI patients treated with dopamine has been above (13, Decreased levels of serum of thyroid hormone e.g. have also been in NTI Some studies have the effects of thyroid hormone in NTI. with T4 was not This may be explained by diminished conversion of T4 to metabolically more active T3 in NTI. For the same treatment with T4 may not be even in NTI patients with low free T4(D) values. Studies evaluating treatment of NTI patients with T3 have either or T3 treatment of patients undergoing or procedures measured by decreased systemic for for or and of These have been related to of metabolism in increase in increase in high and increase in uptake of in the plasma There has been evidence of increased from of T3 the observed effects of T3 in the above-mentioned studies were or is not known. It is however, that short term treatment with of T3 was associated with several effects in NTIS. more should be about the patient population, and any effects of treatment of the NTIS with T3. Alterations in thyroid function tests are very common in patients with NTI. and are in these of these abnormalities is to in the diagnosis of thyroid disease. of serum TSH, free T4, and free T3 levels by direct equilibrium dialysis/RIA methods yield most information in the setting of NTI. with low free T4 by these methods and normal or low TSH have secondary hypothyroidism. This may be due to NTI per administered for treatment of NTI, or associated pituitary or hypothalamic the may require of cortisol and/or A serum TSH level above μU/mL primary findings of low free T4, and positive antithyroid antibodies help establish the An elevated serum concentration of rT3 argues against hypothyroidism. Studies have of treatment of NTI patients with T4. Some studies have a of treatment with T3 in cases, but much more to be There is evidence of by treatment of NTI patients with to of T3. As some NTI patients may indeed be the term should be replaced with NTIS.