Effects of Nonthyroidal Illness on Thyroid Function

Symposium on Thyroid Disease

Effects of Nonthyroidal Illness on Thyroid Function ]oseph M. Tibaldi, M.D.,* and Martin I. Surks, M.D.t

The first detailed analysis of parameters of thyroid function in patients with nonthyroidal non thyroidal illness was the report of Oppenheimer et al., 34 which documented a decrease in serum thyroxine-binding prealbumin, plasma protein binding of thyroxine (T4), and either normal or decreased serum protein-bound iodine (PBI) in hospitalized sick patients who did not have thyroid disease. These initial findings have been amplified and extended by many investigators during the last two decades, facilitated by radioimmunoassay methodology for the measurement of thyroid hormones and thyrotropin (TSH) and the availability of thyrotropin-releasing hormone (TRH), for use in clinical investigation. Because of the ready availability of accurate and sensitive measurements of these hormones, it has become clear that nonthyroidal non thyroidal illness may be associated with a diversity of effects on thyroid hormone homeostasis. Thus serum T4 concentration may be elevated, nor3' -triiodothyronine (T3) values are usually demal, or low. Serum 3, 5, 3'-triiodothyronine creased, and serum TSH is generally normal or minimally elevated. Knowledge of these changes is essential to enable the clinician to decide whether altered thyroid hormone levels are due to the nonthyroidal non thyroidal disease alone or to associated thyroid dysfunction. In addition, the apparent dissociation between serum thyroid levels that are consistent with hypothyroidism and the euthyroid clinical state of the patient has raised many questions concerning thyroid hormone action in nonthyroidal non thyroidal disease as well as the question of thyroid hormone treatment to restore hormone levels to the normal range. This article will focus on these biologic questions and provide a framework for interpretation of various measurements of thyroid function. More detailed citations may be found in several recent reviews. 13, 49 PhYSician, De**Fellow, Division of Endocrinology and Metabolism, and Assistant Attending Physician, partment of Medicine, Montefiore Medical Center; and Instructor in Medicine, Albert Einstein College of Medicine, Bronx, New York tHead, Division of Endocrinology and Metabolism; Attending Physician; and Professor of Medicine, Montefiore Medical Center and the Albert Einstein College of Medicine, Bronx, New York (CAI6463-10 and U. S. Public Health Service (CA16463-10 Supported in part by grants from the V.S. CA24608-05).

America-Vo!' 69, No. 5, September 1985 Medical Clinics of North America-Vol.

899

900

JOSEPH ]OSEPH

M.

A~D MARTIN TIBALDI AND

I.

SURKS

LOW SERUM T3 The thyroid gland secretion is predominantly T4; only about 20 per 22. 39 Moreover about cent of the T3 pool is derived from the thyroid gland. 22, 80 per cent of the extrathyroidal T3 pool is produced from T4 by monodeiodinatIon in peripheral tissues. Most of the biologic effects of the thydeiodinatlon roid hornlones hormones are due to T3 and are initiated by an interaction between 31 With this physiologic backT3 and the nuclear thyroid hormone receptor. 31 11. 17,46 17. 46 have reported ground, it is surprising that a number of investigators 33,. 11, that serum T3 is decreased in patients with nonthyroidal non thyroidal disease who had clinical ,evaluation. been considered euthyroid on clinical· evaluation. In a representative study carried out at Montefiore Medical Center, Bermudez et al. 3 measured serum thyroid hormone and 1'SH TSH levels in patients who were hospitalized for non thyroidal disorders and cared for by the internal a variety of different nonthyroidal medicine service. Although all patients were clinically euthyroid, 70 per cent had serum T3 values in the hypothyroid range, less than 90 ng per dl. The mean serum T3 of the sick patients, 78.4 ± ± 38.3 (SD) ng per dl was significantly less than that of control patients, 134.0 ± 29.3 ng per dl. normalize even after corMoreover, most of the low serum values did not norlnalize rection for changes in hormone binding by serum binding proteins. Serum T4 levels were comparable to those of control patients, and serum TSH was not significantly increased in the same group of patients. Thus, more than one half of randomly selected patients hospitalized on a general internal medical service may have decreased serum T3. Nevertheles, these patients did not appear hypothyroid by clinical criteria or, for the most part, by other laboratory measures. The cause of the decrease in serum T3 in nonthyroidal disease is now known to be a decrease in production of T3 from nooT4. The activity of the 5' -deiodinase enzyme, appears diminished in nonthyroidal illness as well as during voluntary starvation, an experimental model which simulates the "low T3 syndrome." Changes that occur during starvation will be discussed below. Since the sarne same enzyme seems to be responsible for deiodination of both T4 and 3,3' ,5' -triiodothyronine (reverse. non thyroidal illness is associated with an increased level of this T3, rT3) , 14 nonthyroidal biologically inactive iodothyronine, which is due to a decrease in its rate of metabolism. Since serum T3 is significantly decreased in a large proportion of patients with nonthyroidal disease, measurements of T3 are of limited value in the diagnosis of thyroid dysfunction in patients with significant nonthyroidal disorders.

CHANGES IN SERUM T4 Oppenheimer et al. 34 were among the first to report that serum T4 (then measured as serum PBI) was decreased to below the normal range in about 50 per cent of patients with nonthyroidal disease. Low serum T4 in sick patients has been noted by many groups, particularly in critically ill patients or in patients with chronic liver disease. Despite the depressed levels of both T4 and T3 in extremely ill patients, the clinical evaluation is

EFFECTS OF NONTHYROIDAL ILLNESS ON THYROID FUNCTION

901

generally compatible with the euthyroid state and serum TSH is not ele40 measured thyroid hormone levels in 86 patients vated. Thus, Slag et al. 40 who were hospitalized in an intensive care unit. Twenty-two per cent had decreased serum T4 and normal TSH. A strong correlation was noted between mortality and the decrease in serum T4. Eighty-four per cent of patients with a serum T4 less than 3 J.1g f,Lg per dl died; the mortality rate was J.1g per dl 50 per cent for patients with a serum T4 level between 3 and 5 f,Lg and was 15 per cent for those with a serum T4 greater than 5 J.1g f,Lg per dl. Only two of these 86 patients had primary hypothyroidism and both of these individuals had an increase in serum TSH. Thus, a decrease in serum T4 alone should not be considered proof of primary hypothyroidism in patients with nonthyroidal non thyroidal disease unless the serum TSH is elevated as well. A further complication in the diagnosis of thyroid disease in patients with nonthyroidal non thyroidal illness is the fact that some individuals may have an ele20 reported on 15 patients who had an vation in serum T4. T4.4.4, 5, 20 20 Gavin et al. 20 increase in serum T4 and free T4 index on admission to the hospital. Specific treatment was administered only for the underlying non thyroidal disnonthyroidal order. During recovery, the serum T4 concentration decreased to the normal range and the serum T3, initially low, also returned to normal in most TRH was normal in six of these patients patients. The TSH response to TRI-I during the time when serum T4 was elevated. Thus, the increase in serum T4 was not a reflection of hyperthyroidism in these patients. Similar find43 Howings have been reported in patients with acute psychiatric illness. 55,, 43 ever, in these individuals, the TSH response to TRH is often blunted during the acute psychosis. psychosis, Nevertheless, as with the other nonpsychiatric diseases, serum hormone values and TRH responsiveness return to normal when the acute psychiatric syndrome is improved. These considerations suggest that it may be difficult to decide whether an elevated serum T4 value indicates associated hyperthyroidism in patients with nonthyroidal disease. Physical examination may be helpful in young patients but may not be rewarding in the elderly because of the high incidence of apathetic hyperthyroidism which occurs in this group. A normal value or decrease in serum T3 should not in itself exclude the diagnosis of hyperthyroidism. Several reports indicate that hyperthyroid patients with associated nonthyroidal non thyroidal disease may have normal or rarely low serum T3 values; serum T3 may then increase to high values after successful management of the associated nonthyroidal non thyroidal disease. 44,, 18 An increase in the thyroidal uptake of radioiodine will support the diagnosis of hyperthyroidism and a normal TSH response to TRH will support the impression of euthyroidism. However, as discussed below, TSH regulation may not be completely normal in nonthyroidal non thyroidal disease. Because of these considerations, the decision to treat patients with nonthyroidal disease because of an elevation in serum T4 must be tempered. It is probable that transient abnormalities in T4 concentration are much more common than intrinsic thyroid disease in patients with a high T4 on admission to the hospital. Thus, it seems prudent to follow such individuals in order to evaluate the thyroid state after resolution of the nonthyroidal illness.

902 902

M. TIBALDI AND MARTIN JOSEPH M.

I. SURKS

FREE THYROID HORMONES Patients with nonthyroidal illness almost invariably exhibit a decrease in the binding of T4 and T3 to the serum binding proteins and therefore, an increase in the percentage of serum thyroid hormones in the free (unbound) form. This is best measured by an equilibrium dialysis method,34 by which the percentage of unbound hormone is designated the dialyzable fraction or free fraction. The free fraction or dialyzable fraction is almost always increased in patients with nonthyroidal disease. Initial investigations into the mechanisms underlying this change indicated a rapid and marked decrease in concentration of serum thyroxine-binding prealbumin (TBPA) in nonthyroidal non thyroidal disease, probably due to a decrease in synthesis of the pro34,,42 tein. 34 42 Secondly, more recent reports suggest there is also an increase in the proportion of desialated forms of thyroxine-binding globulin (TBG) in the serum of patients with nonthyroidal non thyroidal disease and that this altered bind37 ing protein has a decrease in binding affinity for thyroid hormones. 37 Both the decreased binding affinity of TBG and decreased concentration of TPBA contribute to a decrease in serum binding of thyroid hormone in patients with nonthyroidal non thyroidal disease. However, from a quantitative point of view, these changes in concentration and binding affinity of the thyroxine-binding proteins may not completely account for the increase in free fraction or dialyzable fraction that is observed. 22,, 53 Although the T4 binding non thyroidal illness, illness,22 measurement capacity of TBG is clearly diminished in nonthyroidal of TBG protein by radioimmunoassay reveals either normal or only moder12 postulated the ately low values. 16 Because of these findings, Chopra et al. 12 existence of an inhibitor of thyroid hormone binding to serum binding proteins in these conditions. This was demonstrated experimentally by adding serum of sick patients to pooled normal serum. The dialyzable fraction 'of the pooled normal serum was significantly increased by this addition. Chopra et al. 15 further suggested that the inhibitor of thyroid hormone binding might be a tissue component that leaks into the circulation when there is compromise in tissue integrity. In support of this hypothesis is the finding that addition of homogenates of several human or rat tissues to normal se- . rum increased the dialyzable fraction. Further evidence for the presence of circulating inhibitors of hormone non thyroidal disease was presented by Oppenheimer et al. 33 binding in nonthyroidal They noted a discordance between estimates of free thyroid hormones using charcoal uptake and equilibrium dialysis. The authors postulated that a nondialyzable factor that inhibited the binding of thyroid hormones to serum proteins and charcoal was present in the serum of 74 per cent of sick patients. Since the inhibitor of binding also affected the binding of iodothyronines both by serum proteins and the inert substances such as charcoal or resin, it seems that the resin uptake is not the optimal method for assessing in vivo changes in binding of thyroid hormones or free thyroid hormone concentrations in nonthyroidal disease. Oppenheimer et al. 33 also reported evidence that the binding inhibitor diminished cellular uptake of thyroid hormones as well as the binding of thyroid hormones by serum proteins. This is in agreement with the findings of Pardridge et al. 36 who showed that the amount of circulating T4 or

l\"ONTHYROIDAL ILLNESS ON THYROID FUNCTION EFFECTS OF NONTHYROIDAL

903

T3 in serum of sick patients that was available for transport into the liver was decreased in proportion to serum total hormone levels and was not related to the free T4. Thus, the cellular uptake of thyroid hormones appears to be influenced by many factors including the magnitude of hormone binding by cellular constituents and serum binding proteins and the presence hormone binding by serum and cells. Much work of inhibitors that affect horlTIone remains to be done for a more complete understanding of the mechanism underlying the exchange of thyroid horlllone hormone between serum and cells and its possible regulatory role in the metabolic activity of thyroid hormones.

FREE THYROID HORMONE ASSAYS IN NONTHYROIDAL ILLNESS The equilibrium dialysis technique remains the standard test for the measurement of free T4. However, it is a time-consuming and, therefore, expensive test. Circulating binding inhibitors result in augmented values for free T4 measured by equilibrium dialysis in comparison with methods employing resins, charcoal, or other nonspecific binders of thyroid hormones. In nonthyroidal illness, the free T4 concentration measured by equilibrium dialysis is frequently normal, 16, 24 whereas the less expensive assays of free T4 index (FT4-I), which may utilize resins, may result in values below the normal range. For most patients with a low FT4-I, the absence of an increase in TSH should exclude the diagnosis of primary thyroid failure. In such patients, the FT4 by equilibrium dialysis could be done to exclude the much less common diagnosis of secondary hypothybv clinical circumstances. It is notable that a number roidism if indicated by ac~te nonthyroidal disease may have a normal serum T4 of patients with an acute and an increase in the dialyzable fraction, which results in an increase in 34 At the present time, no biological consequences have been attribFT4. 34 uted to the increase in FT4 in these patients. However, it is unlikely that the increase in FT4 is responsible for the maintenance of the euthyroid state in patients with nonthyroidal non thyroidal disease and decreased serum T3 since an inverse correlation between free T4 and serum T3 has not been found. 33 Thus, the increase in FT4 probably does not provide an increment in biologic activity, which offsets the decrement in biologic activity anticipated from a decrease in serum T3. Many rapid assays for FT4 measurement are commercially available. Kaptein et al. 24 found that FT4 values using the Clinical Assays and Abbott Laboratories kit gave results comparable to those with the equilibrium dialysis method. The results of five other commercially available assays gave lower values. However, Melmed et al. 29 and Slag et al. 41 found that results of all methods, including equilibrium dialysis, resulted in considerable overlap with the hypothyroid range and gave nonspecific results in severely ill patients. Although the results of these studies are somewhat different, patient P9pulations of the intensive care units studied do appear comparable. Until these data are extended to provide better guidance for the selection of methods for FT4 determination, all such tests should be interpreted cautiously in critically ill patients.

904

]OSEPH JOSEPH M. TIBALDI AND MARTIN

I. SURKS

THYROID HORMONE METABOLISM To determine the mechanism for the changes in serum thyroid hormone concentrations, many investigators have examined the production and turnover rates of thyroxine in nonthyroidal disease. Bellabarba et al. 2 reported that the serum half-life of T4 of severely ill patients was shorter than that of controls (5.9 ± 0.9 days vs. 7.7 ± 0.2 days). Similarly, 52 found a shortened half-life for thyroxine in monkeys infected with Woeber52 Diplococcus pneumoniae, and Kaptein et al. 23 reported an increase in metabolic clearance rate of T4 in patients with critical illness. Stockigt et al., 44 in a prospective study, made frequent serum T4 determinations in three patients admitted to the intensive care unit. They noted a half-life for T4 equivalent to 12 to 24 hours. In this study, there was a 20 per cent decrease in serum thyroxine-binding globulin as well. Thus, these studies suggested an increase in the clearance of T4 as well as the presence of an inhibitor or T4 binding. Kaptein et al. 25 performed a more detailed analysis of thyroid hormone ,..,g per dl and kinetics in 16 sick patients with a serum T4 of less than 3.0 J.Lg a normal TSH. A two-fold increase in the metabolic clearance rate of T4 was observed in patients with nonthyroidal illness. This change was quantitatively similar to that of euthyroid subjects who have a decrease in serum thyroxine-binding globulin. This suggests that the increase in T4 clearance is secondary to an increase in the free T4 fraction. An increase in the metabolic clearance rate of T3 was also noted. Moreover, the rate of exit of T4 from the intravascular space decreased, suggesting an impairment in the extravascular binding of T4. The mechanism for such a phenomenon may involve the circulating inhibitors of cellular and serum protein binding of T4 discussed above. 12, 33 The metabolic clearance rate of reverse T3 was also reduced, consistent with a decrease in activity of 5' -deiodinase in nonthynon thyroidal disease. These findings suggest that the T4 production rate is normal in the majority of patients with nonthyroidal illness who are not treated with dopamine. This conclusion is important when considering the biologic impact of these changes. Patients treated with dopamine have a decreased. 26 have shown that dopamine treatment production rate ofT4. of T4. Kaptein et al. 26 results in a decrease in serum TSH and T4 in normal individuals as well as non thyroidal illness. illness, in patients with nonthyroidal REGULATION OF TSH In addition to changes in thyroid hormone binding by serum proteins and in hormone distribution and metabolism, decreased TSH secretion may also contribute to the low serum T4 and T3 concentration observed in patients with nonthyroidal disease. The stress of illness and hospitalization is often accompanied by increased secretion of adrenocortical steroids and Utige~l showed that increased plasma cortisol concentration. Wilber and Utige~l glucocorticoid administration decreases TSH secretion by a mechanism that may be independent of TRH. Thus, elevated cortisol production may inhibit TSH secretion in the sick patient. Decreased food ingestion is another

EFFECTS OF NONTHYROIDAL ILLNESS ON THYROID FUNCTION

905 905

factor prevalent in hospitalized patients with nonthyroid non thyroid disease which also may result in decreased TSH secretion. Carlson et al. 8 reported that the TSH response to TRH was blunted in obese men during a prolonged fast. More recently, Borst et al. 6 demonstrated that the TSH response to TRH (60 hours) fastis decreased in hypothyroid patients even during short-term (60 ing. Moreover, the elevated basal serum TSH of several of these hypothyroid patients decreased into the normal range during fasting. These data suggest that the elevated serum TSH of hypothyroid patients who have an associated nonthyroidal disease may be somewhat attenuated because of poor caloric intake. However, a decrease in TSH secretion may depress serum TSH into the normal range in some patients with mild hypothyroidism but generally not in those with moderate or severe thyroid insufficiency. In most laboratories, the limit of sensitivity of serum TSH determinations approximates the lower border of the normal range. Thus, few laboratories are able to measure a decrease in serum TSH concentration. In a prospective study of 35 patients receiving bone marrow transplant, Wehmann et al. 50 used a sensitive TSH assay to show that basal serum TSH decreased in 17 of 19 patients who had a decline in serum T4, T3, and FT4. These findings suggest that decreased TSH secretion contributes to the decrease in serum T4 and T3 in some sick patients. Most reports indicate that the TSH response to TRH is normal in panon thyroidal disease. 13, 19 However, these findings demonstrate tients with nonthyroidal only that the thyrotrope is responsive to a pharmacologic dose of TRH and non thyroidal disease are do not show that the thyrotropes of patients with nonthyroidal normally responsive to small changes in serum T4 and T3 in a manner similar to normal individuals. If the thyrotropes of sick patients are normally responsive to decreased serum T4 and T3 levels, the normal serum TSH which is frequently associated with decreased serum T4 and T3 could be considered strong evidence in favor of the euthyroid state. To explore this issue, we tested the ability of the thyrotropes of sick patients to respond to a small decrease in serum T4 and T3 by measuring the integrated TSH re28 38,48 had sponse to TRH before and after treatment with iodides. 28 Others 38 shown in normal subjects that iodide treatment results in about a 10 per cent decrease in serum T4 and T3, which is associated with a doubling of the TSH response to TRH. Our study of 23 sick patients with a mean age of 65.0 years (range, 20 to 82 years) revealed that TRH-induced TSH secreof65.0 tion was augmented appropriately in response to decreased serum T4 an~ T3 in only one half of the group. The thyrotropes of the remaining patients in this study were apparently not able to respond to a small decrement in serum thyroid levels. Although there was no correlation between the patient's age and responsiveness of TSH secretion, we carried out a similar study in healthy 35 Similar to our findings in sick patients, TSH regulaelderly individuals. 35 tion was not normal in about half the elderly patients. Thus, advanced age may be an additional factor that contributes to abnormal TSH regulation in hospitalized patients. The failure of TSH to rise in these studies of experimental hypothyroidism suggests that a normal serum TSH may not be considered absolute proof of the euthyroid state in sick patients. It therefore appears prudent to rely on clinical evaluation as well as an increase in se-

906

JOSEPH

M.

TIBALDI AND MARTIN

1. I.

SURKS

rum TSH to establish the diagnosis of primary hypothyroidism in such patients. These studies demonstrated abnormal TSH regulation in half the patients with nonthyroidal disease and low serum T3. Conversely, TSH regulation was normal in the remainder. We have further investigated the mechanism underlying the apparent normal TSH regulation of these panon thyroidal disease, the rat bearing a transtients in an animal model of nonthyroidal plantable Walker 256 carcinoma. 27 Similar to some sick patients, these rats have normal TSH regulation even though they also have decreased levels of serum T4 and T3 and decreased free hormone concentrations. Since locally generated T3 contributes a large proportion of the pituitary T3 pool,27 we studied pituitary T4-to-T3 conversion rates and pituitary T3 content in tumor-bearing rats. The rate of T3 production from T4 was increased in tumor-bearing animals, but the magnitude of this increase was not sufficient to offset the marked depletion of the extrathyroidal T4 pool. Thus, despite an increase in the rate of T3 generation, pituitary T3 remained significantly decreased in tumor-bearing rats. The finding that TSH regulation remained normal despite a decrease in pituitary nuclear T3 suggests that the setpoint for TSH regulation is shifted toward a lower T3 value. In other words, this pituitary response to T3 is augmented in the sick rat.

IMPACT OF DECREASED SERUM THYROID HORMONE LEVELS Because of numerous confounding variables, measurements of oxygen consumption would be difficult to interpret in patients with nonthyroidal non thyroidal disease. Because fasting of normal individuals produces changes in serum non thyroidal illness, illness,88 various thyroid levels similar to those in patients with nonthyroidal findings in fasted patients may be relevant to the sick patient. Since a continuing supply of glucose is essential for brain function during starvation, the production of substrates for gluconeogenesis from muscle breakdown becomes critical for survival. Thus, if decreased serum T3 results in de-. creased metabolic activity of other tissues, products of muscle protein breakdown could be used to a greater extent for support of brain metabolism. The findings of Gardner et al. 19 suggest that the decreased serum T3 levels of fasted patients is associated with reduced muscle protein breakdown and therefore may be considered an important adaptation to decreased caloric ingestion. These findings are particularly relevant to patients with non nonthyroidal thyroidal illness. In studies of fasted patients, this group19 groupl9 others 7, 99 demonstrated that restoration of decreased T3 to normal or and others7, supraphysiologic levels by ingestion of T3 resulted in increased muscle proofT3 tein breakdown. breakdown, Thus, it is possible that decreased serum T3 during periods of illness or fasting serves to preserve muscle mass and possibly fat stores as well. These considerations and the fact that almost all observers non thyroidal illness appear euthyroid despite deagree that patients with nonthyroidal pressed serum T4 and T3 levels argue strongly against thyroid hormone treatment to restore serum T4 and T3 to the normal range. Indeed, Bacci

EFFECTS OF NONTHYROIDAL ILLNESS ON THYROID FUNCTION

907

et al. aL 11 have shown that sick patients augment TSH secretion during their recovery. This response presumably facilitates more rapid restoration of normal serum T4 and T3 after the necessity for this adaptation is over.

WHY DO PATIENTS WITH NONTHYROIDAL ILLNESS APPEAR EUTHYROID? non thyroidal illness and deMost observers agree that patients with nonthyroidal creased serum T3 and T4 appear euthyroid on clinical evaluation. Since T3 is considered responsible for most of the biologic activity of this hormone system, it is surprising that the decreased serum T3 frequently found in patients with nonthyroidal illnes is not accompanied by symptoms or signs of hypothyroidism. Data from animal models of nonthyroidal disease and fasting shed some light on this paradox. In our studies, rats with transplantable Walker 256 carcinoma had changes in serum thyroid hormone levels similar to those of patients with nonthyroidal illness. Serum T4 and T3 and free hormone levels were depressed. We also demonstrated a decrease in 45 In these sick rats, we studied the rethe hepatic nuclear T3 receptor. 45 sponse to T3 of two hepatic enzymes known to be regulated by thyroid hormones: the mitochondrial enzyme, alpha-glycerophosphate dehydrogenase 21 ,, 47 Alpha-GPD activity is consid(alpha-GPD), and cytosol malic enzyme. 21 ered mainly under the control of thyroid hormone, whereas malic enzyme activity is regulated by thyroid hormone, insulin, glucagon, and nutritional factors. Malic enzyme activity results in the production of NADPH, which is needed for fat synthesis. We infused increasing doses of T3 to tumor-bearing and control rats. T3 infusion, we determined the concentration of nuclear T3 At each level of ofT3 receptors and the amount of T3 that was specifically bound to the nuclear receptors. In comparison with control rats, the rate of appearance of alphaGPD was greatly augmented in tumor-bearing rats at all levels of T3 occupancy of nuclear T3 receptor sites. Thus, alpha-GPD regulation was much more sensitive to T3 in tumor-bearing rats. In contrast, the appearance rate of malic enzyme was decreased in tumor-bearing rats at all levels of occupancy of the nuclear T3 receptor. The dose-response curve for malic enzyme was shifted down and to the right. These findings, which are similar to those in fasted rats as reported by Oppenheimer and Schwartz,32 suggest that post-receptor factors modulate specific cellular responses to thyroid hormone; specific responses may become either more sensitive or less sensitive to prevailing levels of thyroid hormones. Thus, in patients with nonthyroidal illness, enhancement of those biologic responses associated with the clinical manifestations of the euthyroid state may account for the normal clinical findings. Since the response of any specific biologic parameter may be either enhanced or depressed by cellular factors that affect postreceptor mechanisms, it is probable that no single biochemical measurement will provide guidance to the thyroidal state of the entire individual in 30 a nonthyroidal disease. 30

908

JOSEPH M. TIBALDI AND MARTIN 1. I. SURKS

SPECIAL MANAGEMENT ISSUES

A particularly difficult problem is how to distinguish very sick euthyroid patiepts with low serum T4 and TSH from patients with hypopituitarism. The history should elucidate predisposing factors such as pituitary surgery, impotence, hair loss, and menstrual abnormalities, and the physical examination should demonstrate abnormal findings such as delayed relaxation time of deep tendon reflexes, dry skin, and testicular atrophy. Further endocrine testing, particularly of the pituitary-adrenocortical, thyroidal, and gonadal axes, should confirm the diagnosis. Of practical concern is the fact that these endocrine test results are generally not available for at least several days, and frequently a decision to institute therapy must be made quickly. In these cases, treatment with glucocorticoids and thyroxine should be instituted according to the usual therapeutic guidelines that clinicians apply to patients who do not have an associated nonthyroidal disorder. Glucocorticoids should be given at least until results of specific tests of adrenocortical function are available. SUMMARY

Despite the absence of thyroid disease, patients with non nonthyroidal thyroidal illness frequently have changes in serum thyroid hormone measurements that may suggest either hypothyroidism or hyperthyroidism. Serum T3 levels are frequently decreased mainly because of a decrease in the rate of T3 production from T4. The free T3 concentration may be either normal or reduced as well. The binding of T4 and T3 by the serum-binding proteins is almost always impaired, resulting in an increase in the dialyzable fraction (free) fraction. This is due to a decrease in the concentration of thyroxinebinding proteins and the presence of circulating inhibitors of binding. If serum T4 concentration remains within the normal range, the free T4 concentration can be increased. However, serum T4 is frequently decreased in patients with chronic and/or severe illness. The decrease in serum T4 in these patients, combined with an increase in the dialyzable fraction, results in normal free T4. In patients who are critically ill, none of the available methods for measurement of free T4 may give results that accurately reflect the euthyroid state. Since T3 is the major active thyroid hormone, it is surprising that patients with decreased serum T3 do not appear hypothyroid. The decrease non thyroidal illness, which in serum T3 is probably an adaptive change to nonthyroidal at least enables the sick patient to conserve protein. The clinical impression of euthyroidism is supported by the finding of a normal serum TSH level in most patients. Although TSH regulation may not be entirely normal in patients with nonthyroidal disease, it is likely that serum TSH will be increased in most sick patients who also have significant thyroid failure. The normal clinical findings in patients with decreased serum T3 may result from an augmentation of those biologic responses associated with the clinical manifestations of the euthyroid state. Several animal models of nonthynon thyroidal disease or starvation show that cells have the ability to modulate

EFFECTS OF NONTHYROIDAL ILLNESS ON THYROID FUNCTION

909

some biologic responses to thyroid hormone. Further study should elucidate the mechanisms underlying these changes. This article has emphasized that no single laboratory measurement may reliably predict the thyroid state in patients with nonthyroidal disease. This fact emphasizes the need for careful clinical evaluation of these patients and judicious use of laboratory tests. Because the changes in thyroid hormone metabolism that occur in nonthyroidal non thyroidal disease probably represent adaptive changes to the illness, treatment with L-thyroxine to restore serum thyroid concentrations to the normal range is not indicated.

REFERENCES 1. Bacci, V., Schussler, G. C., and Kaplan, T. B.: The relationship between serum triiodothyronine and thyrotropin during systemic illness. J. Clin. Endocrinol. Metab., Metab. , 54:122~2235, 54: 1229-2235, 1982. 2. Bellabarba, D., Inada, M., Varsano-Aharon, N., et al.: Thyroxine transport and turnover in major nonthyroidal non thyroidal illness. J. Clin. Endocrinol. Metab., 28: 1023-1030, 1968. 3. Bermudez, F., Surks, M. 1., I., and Oppenheimer, J. H.: High incidence of decreased senon thyroidal disease. J. Clin. Enrum triiodothyronine concentration in patients with nonthyroidal docrinol. Metab., 41:27, 1975. 4. Birkhauser, M., Busset, R, R., Burer, T., et al.: Diagnosis of hyperthyroxinemia when serum thyroxine alone is raised. Lancet, 2:53-56, 1977. 5. Borst, G. C., EH, C., and Burman, F. D.: Euthyroid hyperthyroxinemia. Ann. Intern. Med., 98:366-378, 98:361h378, 1983. 6. Borst, G. C., Osburne, R. C., O'Brien, J. T., et al.: Fasting decreases thyrotropin responsiveness to thyrotropin-releasing ho'rmone: hormone: A potential cause of misinterpretation of thyroid function tests in the critically ill. J. Clin. Endocrinol. Metab., 57:380-384, 57:380--384, 1983. of T3 and reverse 7. Burman, K. D., Wartofsky, L., Dinterman, R. E., et al.: The effect ofT3 T3 administration on muscle protein catabolism during fasting as measured by 3methylhistidine excretion. Metabolism, 28:805-813, 1979. 1. J., et al.: Alterations in basal and TRH-stimu8. Carlson, H. E., Drenick, E. J., Chopra, I. lated serum levels of thyrotropin, prolactin, and thyrOid thyroid hormones in starved obese men. J. Clin. Endocrinol. Metab., 45:707-713, 1977. 9. Carter, W. J., Shakir, K. M., Hodges, S., et al.: Effect of thyroid hormones on metabolic adaptation to fasting. Metabolism, 24:1177-1183, 1975. 10. Cheron, R. R G., Kaplan, M. M., and Larsen, P. R.: R: Physiological and pharmacological influences on thyroxine to 3,5,3'-triiodothyronine conversion and nuclear 3,5,3'64:1402--1414, 1979. triiodothyronine binding in rat anterior pituitary. J. Clin. Invest., 64:1402-1414, 11. Chopra, I. J., Chopra, D., R., et al.: Reciprocal changes in serum concentraU., Smith, S. R, ,5' -triiodothyronine (reverse T3) and 3,3' ,5' -triiodothyronine (T3) in severe tions of 3,3' ,5'-triiodothyronine ,5'-triiodothyronine 41:1043-1049, :1043-1049, 1975. systemic illnesses. J. Clin. Endocrinol. Metab., 41 12. Chopra, I. 1. J., Chua Teco, G. N., Nguyen, A. H., et al.: In search of an inhibitor of thythyroid illnesses. J. Clin. Endocrinol. roid hormone binding to se'rum proteins in non nonthyroid Metab., 49:63-69, 1979. 13. Chopra, I. 1. J., Hershman, J. M., Pardrige, M. D., et al.: ThyrOid Thyroid function in nonthyroidal illnesses. Ann. Intern. Med., 98:946-957, 1983. 14. Chopra, I. J., Solomon, D. H., Chopra, D., U., et al.: Pathways of metabolism of thyroid hormones. Recent Prog. Harm. Horm. Res., 34:521-567, 1978. 15. Chopra, 1. I. J., Solomon, D. H., Chu-Teso, G. N., et al.: An inhibitor of the binding of thyroid hormones to serum proteins is present in extrathyroidal tissues. Science, 215:407-409, 1982. 16. Chopra, I. 1. J., Solomon, D. H., Hepner, G. W., et al.: Misleading low free thyroxine index and usefulnes of reverse triiodothyronine measurement in non-thyroidal illnesses. Ann. Intern. Med., 90:905-912, 1979. 17. Czernichow, P., Dauzet, M. C., Broyer, M., et al.: Abnormal TSH, PRL and GH re-

910 18. 19. 20. 21. 22. 23. 24.

25. 26. 27.

28.

29.

30. 31. 32. 33.

34. 35. 36.

37.

JOSEPH M. M. TIBALDI AND MARTIN

I. SURKS

sponse to TSH releasing factor in chronic renal failure. J. Clin. Endocrinol. Metab., 43:630-637, 1976. a!.: Hyperthyroidism without Engler, D., Donaldson, E. B., Stockight, J. R., et al.: triiodothyronine excess: An effect of severe non-thyroidal illness. J. Clin. Endocrinol. Metab., 46:77-82, 1978. a!.: Effect of tri-iodothyronine reGardner, D. F., Kaplan, M. M., Stanley, C. A., et al.: placement on the metabolic and pituitary responses to starvation. N. Engl. J. Med., 300:579-584, 1979. Gavin, L. A., Rosenthal, M., and Cavalieri, R. R.: The diagnostic dilemma of isolated hyperthyroxinemia in acute illness. J.A.M.A., 242:251-253, 1979. Grajower, M. M., and Surks, M. 1.: Effect of decreased hepatic nuclear L-triiodothyronine receptors on the response of hepatic enzymes to L-triiodothyronine in tumorbearing rats. Endocrinology, 104:697-703, 1979. Ingbar, S. H., and Braverman, L. E.: Active form of thyroid hormone. Annu. Rev. Med., 26:443-449, 1975. Kaptein, E. M., Grieb, D. A., Spencer, C. A., et al.: Thyroxine metabolism in the low thyroxine state of critical nonthyroidal illnesses. J. Clin. Endocrinol. Metab., 53:764771, 1981. Kaptein, E. M., MacIntyre, S. S., Weiner, J. M., et al.: Free thyroxine estimates in nonthyroidal illness: Comparison of eight methods. J. Clin. Endocrinol. Metab., 52:10731077, 1981. Kaptein, E. M., Robinson, W. J., Grieb, D. A., et al.: a!.: Peripheral serum thyroxine, triiodothyronine and reverse triiodothyronine kinetics in the low thyroxine state of thyroidal illnesses. J. Clin. Invest., 69:52&-535, 69:526-535, 1982. acute non nonthyroidal Kaptein, E. M., Spencer, C. A., Kamiel, M. B., et al.: Prolonged dopamine administration and thyroid hormone economy in normal and critically ill subjects. J. Clin. Endo51:387-393, crinol. Metab., 51 :387-393, 1980. Kumara-Siri, M. H., Lee, K. Y., and Surks, M. I.: Regulation of thyrotropin release in tumor-bearing rats. Endocrinology, 109:1760-1768, 109: 1760-1768, 1981. a!.: Variable thyrotropin response to thyroMaturlo, S. J., Rosenbaum, R. L., Pan, C., et al.: tropin-releasing hormone following small decreases in plasma free thyroid hormone concentrations in patients with nonthyroidal diseases. J. Clin. Invest., 66:451--456, 66:451-456, 1980. Melmed, S., Geola, F. L., Reed, A. W., et al.: a!.: A comparison of methods for assessing thyroid function in nonthyroidal illness. J. Clin. Endocrinol. Metab., 54:300-306, 1982. Oppenheimer, J. H.: Thyroid function tests in nonthyroidal disease. J. Chronic Dis., 35:697-701. 1982. phy- . Oppenheimer, J. H., and Dillman, W. H.: Nuclear receptors for triiodothyronine: A phy-' sioloical perspective. In O'Malley, B. W., and Birnhaumer, L. (eds.): Receptors and Vo!. 3, p. 1. Hormone Action. New York, Academic Press, 1984, Vol. Oppenheimer, J. H., and Schwartz, H. L.: Factors determining the level of activity of 3,5,3' -triiodothyronine-responsive hepatic enzymes in the starved rat. Endocrinology, 107:1460-1468, 1980. Oppenheimer, J. H., Schwartz, H. L., Mariash, C. N., et al.: Evidence for a factor in the sera of patients with nonthyroidal disease which inhibits iodothyronine binding by Endocrino!. Metab., 54:757solid matrices, serum proteins, and hepatocytes. J. Clin. Endocrinol. 766, 1982. Oppenheimer, J. H., Squef, R., Surks, M. 1., I., et al.: Binding of thyroxine by serum proteins evaluated by equilibrium dialysis and electrophoretic techniques. Alterations in nonthyroidal disease. J. Clin. Invest., 42: 1769, 1963. 42:1769, Ordene, K. W., Pan, C., Barzel, U. S., et al.: Variable thyrotropin response to thyrotropin releasing hormone after small decreases in plasma thyroid hormone concentrations in patients of advanced age. Metabolism., 32:881-88, 1983. Pardridge, W. M., Slag, M. F., Morley, J. E., et al.: a!.: Hepatic bioavailability of serum thyroid hormones in nonthyroidal illness. J. Clin. Endocrinol. Metab., 53:913-916, 1981. Wellby, M. L.: Slow thyroxine-binding globulin in the pathogenesis of Reilly, C. P., and WeBby, non thyroidal illness. J. Clin. Endocrinol. increased dialysable fraction of thyroxine in nonthyroidal Metab., 57:15-18, 1983.

EFFECTS OF NONTHYROIDAL ILLNESS ON THYROID FUNCTION

911

38. Saberi, M., and Utiger, R. D.: Augmentation of thyrotropin response to thyrotropin-releasing hormone follow small decreases in serum thyroid hormone concentrations. J. 40:43&-441, 1975. Clin. Endocrinol. Metab., 40:435-441, 39. Schimmel, M., and Utiger, R. D.: Thyroidal and peripheral production of thyroid hormones: Review of recent findings and their clinical implications. Ann. Intern. Med., 87:760-768, 1977. 40. Slag, M. F., Morley, J. E., Elson, M. K., et al.: Hypothyroxinemia in critically ill patients as a predictor of high mortality. J. A. M. A., 245:4~5, 245:43--45, 1981. M. K., et al.: Free thyroxine levels in critically ill 41. Slag, M. F., Morley, J. E., Elson, ~1. patients. A comparison of currently available assays. J. A. M. A., 246:2702-2706, 1981. 131 of F31 human 42. Socolow, E. L., Woeber, K. A., Purdy, R. H., et al.: Preparation ofI labeled hunlan serum prealbumin and its metabolism in normal and sick patients. J. Clin. Invest., 44:1600-1609, 1965. 43. Spratt, D. I., Pont, A., Miller, M. B., et al.: Hyperthyroxinemia in patients with acute psychiatric disorders. Am. J. Med., 73:41-48, 1982. R, Barlow, J. W., Lim, C.-F., et al.: Rapid decline in serum T4 during se44. Stockigt, J. R., vere nonthyroidal illness with altered relationship between immunoreactivity and binding capacity of thyroxine binding globulin [Abstract]. In Proceedings of the 59th meeting of American Thyroid Association, 1983. 45. Surks, M. I., Grajower, M. M., Tai, M., et al.: Decreased hepatic nuclear Ltriiodothyronine receptors in rats and mice bearing transplantable neoplasms. Endocrinology, 103:2234-2239, 1978. 46. Talwar, K. K., Sawhney, R. C., and Rastogi, G. K.: Serum levels of thyrotropin, thyroid hormones and their response to thyrotropin releasing hormone in infective febrile illEndocrinol. Metab., 44:398-403, 1977. nesses. J. Clin. Endocrinol.~etab., 47. Tibaldi, J., and Surks, M. I.: Response of hepatic mitochondrial a-glycerophosphate dehydrogenase and malic enzyme to constant infusion of L-triiodothyronine in rats bearing the Walker 256 carcinoma: Evidence for divergent postreceptor regulation of the thyroid hormone response. J. Clin. Invest. 74:705-714, 1984. 48. Vagenakis, A. G., Rapoport, B., Azizi, F., et al.: Hyperresponse to thyrotropin-releasing J. hormone accompanying small decreases in serum thyroid hormone concentrations. J. Clin. Invest., 54:913-918, 1974. 49. Wartofsky, L., and Burman, K. D.: Alterations in thyroid function in patients with systemic illness: the "euthyroid sick syndrome." Endocr. Rev., 3:164-217, 1982. 50. Wehmann, R. E., Gregerman, R. I., Burns, W. H., et al.: Suppression of thyrotropin in RE., RI., the low-thyroxine state of severe nonthyroidal illness. N. Engl. J. Med., 312:546-552, 1985. J. F., and Utiger, R. R D.: The effect of glucocorticoids on thyrotropin secretion. 51. Wilber, J. J. Clin. Invest., 48:2096-2103, 1969. 52. Woeber, K. A.: Alterations in thyroid hormone economy during acute infection with 50:378-387, 1971. Diplococcus pneumoniae in the rhesus monkey. J. Clin. Invest., 50:378--387, 53. Woeber, K. A., and Ingbar, S. H.: The contribution of thyroxine-binding prealbumin to the binding of thyroxine in human serum, as assessed by immunoadsorption. J. Clin. Invest., 47: 1710-1718, 1968. 47:1710-1718, (Dr. Surks) Department of Medicine Montefiore Medical Center III East 210th Street 111 Bronx, New York 10467