Thyroid function in nonthyroidal illness

INTRODUCTION — Assessment of thyroid function in patients with nonthyroidal illness is difficult, especially among those hospitalized in an intensive care unit (ICU). Many of them have low serum concentrations of both thyroxine (T4) and triiodothyronine (T3), and their serum thyroid-stimulating hormone (TSH) concentration also may be low. Previously, these patients were thought to be euthyroid, and the term euthyroid-sick syndrome was used to describe the laboratory abnormalities. However, there is evidence that some of these patients may have acquired transient central hypothyroidism [1,2]. It is possible that the early changes in thyroid function (low T3 and elevated reverse [rT3]) during severe illness are protective in that they prevent excessive tissue catabolism [3]. On the other hand, the subsequent alterations in thyroid hormone levels may be maladaptive, depending on the clinical circumstances [4].

This topic will review the changes in thyroid hormone and TSH production and metabolism that can occur in patients with nonthyroidal illness. The utility of the different tests to assess thyroid function is discussed separately. (See "Laboratory assessment of thyroid function".)

GENERAL PRINCIPLES — Important general principles are as follows [5]:

●Thyroid function should not be assessed in seriously ill patients unless there is a strong suspicion of thyroid dysfunction, since there are many other factors in acutely or chronically ill euthyroid patients that influence thyroid function tests.

●In hospitalized or ill patients in whom TSH measurement is necessary (eg, a sick patient with paroxysmal atrial fibrillation), TSH assays that have a detection limit of 0.01 mU/L should be used to assess thyroid function.

●When thyroid dysfunction is suspected in critically ill patients, measurement of serum TSH alone is inadequate for the evaluation of thyroid function. In such a patient, TSH and free T4 (and often total T3) are needed to differentiate nonthyroidal illness from a thyroid disorder. However, the diagnosis may still be in doubt.

CHANGES IN THYROID HORMONE TESTS — Many hospitalized/ill patients have patterns of thyroid function that are similar to patients with central hypothyroidism: low or low-normal serum total T4, low T3 concentrations, and low, low-normal, or normal TSH [6]. It is possible that these changes in thyroid function during severe illness are protective in that they prevent excessive tissue catabolism [1,3,4].

TSH — Euthyroid patients with nonthyroidal illness, especially those receiving high-dose glucocorticoids (>20 mg/day prednisone or its equivalent), dopamine, or dobutamine may have low serum TSH (figure 1), although it would be unusual for the TSH to be undetectable in an assay with a detection limit of 0.01 mU/L in the absence of thyrotoxicosis. If a more sensitive TSH assay (detection limit of 0.01 mU/L) is used, the serum TSH values in patients being given these drugs are usually in the range of 0.08 to 0.4 mU/L. These values can easily be distinguished from the suppressed values in patients with hyperthyroidism (<0.01 mU/L). (See 'TSH subnormal' below and 'Drugs' below.)

In addition, cytokines and inflammation lower TSH levels, possibly by reducing hypothalamic thyrotropin-releasing hormone (TRH) production [4,7]. Patients with nonthyroidal illness, similar to those with central hypothyroidism from other causes, have a blunted nocturnal rise in serum TSH concentrations but usually have a normal serum TSH response to TRH [8]. Abnormalities in TSH glycosylation that may decrease TSH bioactivity have been found in patients with nonthyroidal illness [9] and also in patients with central hypothyroidism [10]. (See 'Acquired transient central hypothyroidism' below.)

Some hospitalized patients have transient elevations in serum TSH concentrations (up to 20 mU/L) during recovery from severe nonthyroidal illness. (See 'TSH high' below.)

Serum total and free T4

●Total T4 – From 15 to 20 percent of hospitalized patients and up to 50 percent of patients in intensive care units (ICUs) have low serum T4 concentrations (low T4 syndrome) (figure 1). The concentrations are low because of reductions in the serum concentrations of one or more of the three thyroid hormone-binding proteins: thyroxine-binding globulin (TBG), transthyretin (TTR, or thyroxine-binding prealbumin [TBPA]), and albumin.

Since TBG is the major binding protein, low serum T4 values are likely the result of decreased production of normal TBG or production of TBG that binds T4 poorly because it is abnormally glycosylated or is cleaved in the circulation [11].

●Free T4 – Free T4 levels may be normal, increased, or decreased. All methods for assessing free T4 levels are unreliable in severe critical illness, including free T4 by mass spectrometry [12]; a free T4 by equilibrium dialysis sent to a reference laboratory would be the least likely to provide spurious results. Some experts argue a total T4 is more informative than free T4 measurements in critical illness [13].

Small reductions in binding proteins should not alter serum free T4 index or direct free T4 values, and these values are usually normal in patients whose illness is not severe. However, when the concentrations of binding proteins are very low, the T3 uptake test (and therefore the calculated thyroid hormone-binding ratio) fails to correct for the binding-protein deficiency adequately and the serum free T4 index is low (figure 1) [14]. The effects of nonthyroidal illness and drugs on "direct" free T4 measurements may be method dependent and also result in low values (or occasionally spuriously high values). (See "Laboratory assessment of thyroid function", section on 'Serum free T4 and T3'.)

●Inhibitors of T4 binding – An additional problem occurs in patients with more severe critical illness because some of them have circulating substances that inhibit T4 binding to the binding proteins. The result is a further reduction in serum total T4 concentrations and, frequently, low serum free T4 concentrations [15] and low serum free T4 index values. However, the serum free T4 fraction measured by equilibrium dialysis may be normal or even slightly high in these patients [16]. In one study, as an example, none of 25 patients with nonthyroidal illness had a low serum concentration of free T4 by equilibrium dialysis [17].

Controversy exists as to the cause and importance of inhibitors of T4 binding to its binding proteins in patients with nonthyroidal illness [16]. Some data support the possibility that high serum free fatty acid concentrations inhibit T4 binding to serum proteins [18,19]. Serum free fatty acid, particularly oleic acid, concentrations may be high in critically ill patients [20], and their effect on T4 binding may be increased due to hypoalbuminemia because albumin is the major carrier of free fatty acids in serum [21]. However, this phenomenon may be an in vitro effect of fatty acids released during collection and transport of the serum sample [22]. Inhibitors may also interfere with the T3-resin uptake test by interacting with the solid matrices used in the test [23].

Low serum T3 — The majority of hospitalized patients have low serum T3 concentrations, as do some outpatients who are ill (figure 1). Unlike T4, which is produced solely within the thyroid, 80 percent of circulating T3 is produced by the peripheral 5'-deiodination of T4 to T3, a reaction catalyzed by 5'-monodeiodinases (D1 and D2) in organs such as muscle, liver, and kidney (figure 2). 5'-monodeiodination decreases whenever caloric intake is low and in any nonthyroidal illness, even when mild [24]. Liver and skeletal muscle biopsies obtained within minutes after death from ICU patients demonstrate reduced 5'-monodeiodinase activity and increased 5-monodeiodinase (D3) activity (which converts T4 to reverse T3 [rT3]) [25,26]. Patients with fatal illness have low tissue T4 and T3 concentrations [27,28].

Several mechanisms can contribute to the inhibition of 5'-monodeiodination and therefore to the low serum T3 concentrations in patients with nonthyroidal illness. They are:

●High endogenous serum cortisol concentrations and exogenous glucocorticoid therapy [29].

●Circulating inhibitors of deiodinase activity, such as free (nonesterified) fatty acids [30].

●Treatment with drugs that inhibit 5'-monodeiodinase activity, such as amiodarone and high doses of propranolol.

●Cytokines (such as tumor necrosis factor, interferon alpha, nuclear factor kappa-B [NF-kB], and interleukin-6) [7,31-33].

●Serum samples from patients with nonthyroidal illness impair uptake of T4 into cultured rat hepatocytes, thereby reducing the availability of substrate for conversion to T3 [34].

High reverse T3 — Serum rT3 concentrations are high in patients with nonthyroidal illnesses (figure 1), except in those with renal failure [35,36] and in some with acquired immune deficiency syndrome (AIDS) [37,38]. rT3 is the product of 5-monodeiodination of T4 (D3) (figure 2); D3 is induced in nonthyroidal illness, especially in the setting of hypoxia or ischemia [25,39]. The clearance of rT3 to diiodothyronine (T2) is reduced in nonthyroidal illness because of inhibition of the 5'-monodeiodinase activity [40]. As a result, serum levels of rT3 are elevated. (See "Thyroid hormone synthesis and physiology".)

Acquired transient central hypothyroidism — Patients with severe nonthyroidal illness may have acquired transient central hypothyroidism [1]. This suggestion is supported by the following observations:

●A prospective study evaluated changes in thyroid function in patients undergoing bone marrow transplantation; serum TSH concentrations fell coincident with declines in serum T4 concentrations [41].

●A study of critically ill patients recovering from nonthyroidal illness demonstrated that a rise in serum TSH concentration (which transiently reached supranormal values in some patients) preceded normalization of serum T4 concentrations [42].

●Patients with nonthyroidal illness, similar to those with central hypothyroidism from other causes, have a blunted nocturnal rise in serum TSH concentrations but usually have a normal serum TSH response to TRH [8].

●Abnormalities in TSH glycosylation that may decrease TSH bioactivity have been found in patients with nonthyroidal illness [9] and also in patients with central hypothyroidism [10].

●TRH mRNA in the paraventricular nucleus of the hypothalamus was reduced in patients who died with nonthyroidal illness in one report and was correlated with premortem serum T3 and TSH concentrations [43].

●TRH (protirelin) infusion in patients with critical illness raises serum TSH, T4, and T3 concentrations [44].

●Infusion of interferon alfa to eight normal men caused a fall in serum TSH and T3 concentrations as well as a rise in the serum concentrations of rT3 and interleukin-6, thus mimicking the thyroid metabolic changes of serious illness, except that there was no fall in the serum T4 concentration [7].

●Animal studies support the hypothesis that inflammation (eg, injection of endotoxin in rodents) upregulates hypothalamic deiodinase D2, which increases local T3 production and lowers TRH mRNA expression [4].

DRUGS AND DISORDERS ASSOCIATED WITH NONTHYROIDAL ILLNESS

Drugs — Hospitalized patients frequently receive medications that have important effects on thyroid function or on thyroid function tests (table 1). Dopamine, dobutamine, glucocorticoids, furosemide, nonsteroidal anti-inflammatory drugs (NSAIDs), heparin, anticonvulsants, metformin, and drugs that affect thyroxine-binding globulin (TBG) can alter thyroid function tests in multiple complex ways that are often incompletely understood. (See "Drug interactions with thyroid hormones" and "Euthyroid hyperthyroxinemia and hypothyroxinemia".)

Disorders — Several disease states are associated with abnormal thyroid function tests including acute hepatitis, hepatoma, acute intermittent porphyria, acromegaly, nephrotic syndrome, Cushing syndrome, acute psychosis, and depression. (See "Euthyroid hyperthyroxinemia and hypothyroxinemia" and "Endocrine dysfunction in the nephrotic syndrome".)

Some patients with acute psychiatric illnesses, particularly schizophrenia, have transient elevations in serum T4 concentrations with or without low serum TSH concentrations [45-47]. Patients with severe depression may have changes similar to those of patients with glucocorticoid excess. (See "Euthyroid hyperthyroxinemia and hypothyroxinemia".)

EVALUATION

Biochemical tests — It is important to recognize nonthyroidal illness and to try to differentiate it from an underlying primary thyroid disorder, although the diagnosis may remain in doubt.

We suggest initial measurement of TSH and free T4. Some clinicians also measure total T3 and total T4 at the time of the initial testing to expedite decision making. Measurement of TSH alone can be misleading since a premorbid normal TSH is usually suppressed to subnormal values and a premorbid elevated TSH can rarely be suppressed to normal values. Serum TSH assays that have a detection limit of 0.01 mU/L should be used in assessing thyroid function in critically ill patients [48].

TSH subnormal — Almost all patients who have a subnormal but detectable serum TSH concentration (greater than 0.05 mU/L and less than 0.3 mU/L) will be euthyroid when reassessed after recovery from their illness. In contrast, approximately 75 percent of patients whose TSH is undetectable (<0.01 mU/L) have thyrotoxicosis.

●In patients without a strong clinical suspicion of thyroid disease and minor TSH abnormalities (TSH between 0.05 and 0.3 mU/L with normal or low free T4), we reassess thyroid tests (TSH, free T4) after recovery.

If true central hypothyroidism due to hypothalamic or pituitary disease remains in the differential diagnosis, measurement of serum cortisol can be helpful as it would be elevated in critical illness and low (or inappropriately normal) in patients with true central hypothyroidism. (See "Central hypothyroidism".)

Measurement of serum rT3 is only rarely useful in hospitalized patients to distinguish between nonthyroidal illness (high values) and central hypothyroidism (low values); the values are low in the latter patients because of low production of the substrate (T4) for rT3. In patients with mild hypothyroidism, however, serum rT3 concentrations may be normal or even slightly high [49], limiting its usefulness.

●In patients with suspected hyperthyroidism (as an example, in a sick patient with paroxysmal atrial fibrillation) and TSH <0.05 mU/L with elevated or high normal free T4, we measure total T3 (and sometimes total T4) to help distinguish between hyperthyroidism and nonthyroidal illness.

The serum T3 value should be high (or high normal) in hyperthyroidism, but low (or low normal) in nonthyroidal illness. Rarely, a very sick patient with hyperthyroidism will have a low serum T3 concentration. In critically ill patients with suspected hyperthyroidism (TSH usually <0.01 but can be as high as 0.05 mU/L, and normal or high normal serum T4 and/or T3), we suggest antithyroid drug therapy (thionamides), with a plan for reassessment after recovery from the nonthyroidal illness. (See "Diagnosis of hyperthyroidism", section on 'Critically ill hyperthyroid patients' and "Thionamides in the treatment of Graves' disease", section on 'Initiation of therapy'.)

TSH high — As noted above, some hospitalized patients have transient elevations in serum TSH concentrations (up to 20 mU/L) during recovery from nonthyroidal illness [48]. Few of these patients prove to have hypothyroidism when reevaluated after full recovery from their illness. Patients with serum TSH concentrations over 20 mU/L usually have permanent hypothyroidism [50]. Our approach depends on the degree of TSH elevation and the clinical suspicion for underlying hypothyroidism:

●TSH between upper limit of normal and <10 mU/L – If the patient appears to be recovering from the underlying illness, we repeat the TSH in one to two weeks. Few of these patients prove to have hypothyroidism when reevaluated after recovery from their illness.

●TSH 10 to 20 mU/L – Treatment with levothyroxine may be appropriate depending on the free T4 level, clinical suspicion of hypothyroidism, and degree of nonthyroidal illness. (See "Diagnosis of and screening for hypothyroidism in nonpregnant adults", section on 'Diagnosis'.)

If uncertain, repeat the TSH and free T4 in one to two weeks. Thyroid function tests may improve in patients recovering from nonthyroidal illness.

●TSH ≥20 mU/L – Assess the free T4 level.

•Free T4 low – Hypothyroidism is likely. Initiate thyroid hormone. In the absence of suspected myxedema coma, repletion should be cautious, beginning with approximately half the expected full replacement dose of T4 (levothyroxine). In suspected myxedema coma, or in critically ill patients who cannot ingest or absorb oral medications, thyroid hormone should be given intravenously. (See "Treatment of primary hypothyroidism in adults", section on 'Standard replacement therapy' and "Myxedema coma", section on 'Thyroid hormone'.)

•Free T4 normal – Repeat TSH and free T4 in one to two weeks. Thyroid function tests may improve in patients recovering from nonthyroidal illness.

TSH normal — Most patients with normal TSH and low free T4 will be euthyroid when reassessed after recovery from their illness. Rarely, an elevated TSH in a sick hypothyroid patient will be suppressed into the normal range. We reassess thyroid tests (TSH, free T4) after recovery. If true central hypothyroidism due to hypothalamic or pituitary disease remains in the differential diagnosis, measurement of serum cortisol can be helpful as it would be elevated in critical illness and low (or inappropriately normal) in patients with true central hypothyroidism. (See "Central hypothyroidism".)

MANAGEMENT — In critically ill patients with low free T4 and total T3 who do not appear to have an underlying primary thyroid disorder, we recommend not treating with thyroid hormone. There is no evidence that thyroid hormone replacement is beneficial for patients with critical illness who have low serum T4 or T3 concentrations, or for patients undergoing coronary artery bypass graft surgery (CABG), whose serum T3 concentrations are known to decrease in the perioperative period [3,6]. In addition, during fasting, when there is an associated decrease in serum T3 concentrations that spares muscle protein, T3 (liothyronine) replacement results in increased catabolism with breakdown of skeletal muscle [51,52].

●Critical illness – Thyroid hormone replacement does not appear to be beneficial for critically ill patients with low serum T3 and/or low T4 concentrations. In a randomized trial of burn patients with low free T4 index and free T3 index levels, T3 replacement had no effect on mortality or metabolic rate when compared with placebo [53]. In a second trial, administration of T4 (levothyroxine) to 23 critically ill patients with low serum T4 concentrations did not alter either mortality or outcome [54].

●Cardiac disease – During and after cardiopulmonary bypass, there is a transient decrease in serum T3 concentrations, which may contribute to postoperative hemodynamic problems [52]. While animal data and anecdotal clinical experience had suggested that T3 repletion might improve outcomes after cardiopulmonary bypass [55,56], clinical trials have not demonstrated such a benefit [57-59].

In a systematic review of 14 randomized trials evaluating the administration of T3 in euthyroid adult patients in the immediate postoperative period (13 cardiac surgery, one renal transplantation), intravenous (IV) T3 administration increased cardiac index [60]. Mortality was not affected by high-dose IV T3 and could not be assessed for low-dose IV or oral T3. In one of the trials included in the meta-analysis, 142 patients with coronary heart disease undergoing CABG were randomly assigned to intravenous T3 therapy (0.8 mcg/kg bolus followed by an infusion of 0.113 mcg/kg/hour for six hours) at the completion of surgery or placebo [57]. Although the mean cardiac index was higher and systemic vascular resistance was lower in the T3 group compared with placebo, there were no differences in the incidence of arrhythmia, need for inotropic or vasodilator drugs during the 24 hours after surgery, or in perioperative morbidity and mortality. A subsequent randomized controlled trial using oral T3 (20 mcg every 12 hours) also failed to demonstrate a clinical benefit [61].

N-acetylcysteine is an antioxidant that prevents reductions in glutathione, a cofactor of the D1 deiodinase. In a randomized controlled trial of patients presenting with myocardial infarction, N-acetylcysteine ameliorated the fall in T3 and rise in reverse T3 (rT3) [62]. Whether this is beneficial or harmful has yet to be determined. (See 'Prognosis' below.)

PROGNOSIS — In patients with ischemic heart disease, changes in thyroid tests consistent with nonthyroidal illness are associated with an increase in overall mortality (hazard ratio [HR] 2.61, 95% CI 1.89-3.59) and major adverse cardiac events (HR 2.22, 95% CI 1.71-2.89) [63]. In patients hospitalized with coronavirus disease 2019 (COVID-19), there was an increased risk of death, intensive care unit (ICU) admission, or need for ventilation in patients with severe nonthyroidal illness compared with those with normal thyroid function (HR 23, 95% CI 5.8-93) [64].

The magnitude of the changes in thyroid function in patients with nonthyroidal illness varies with the severity of the illness (figure 1). Low serum T3 has been shown to correlate with increased hospital stay, ICU admission, and need for mechanical ventilation in patients with acute heart failure [65] and predicts 30-day mortality in patients with community-acquired pneumonia [66]. The serum T4 value also correlates with outcome in critically ill patients; values under 3 mcg/dL have been associated with mortality rates in excess of 85 percent [67].

However, there is some evidence that low serum T3 levels may be beneficial in critically ill patients. In a randomized trial of early versus late parenteral nutrition in critically ill ICU patients, tolerating a nutritional deficit for one week was associated with fewer complications and faster recovery from organ failure [68]. The patients who tolerated the nutritional deficit for one week had lower TSH, T4, T3, and T3/reverse T3 (rT3) ratios [69]. Lower T3 but higher T4 was associated with a higher likelihood of early alive ICU discharge. These data suggest that inactivation of T4-to-T3 conversion during starvation and illness may be a beneficial adaptation, while very low T4 levels due to "acquired central hypothyroidism" in more critically ill patients are associated with deleterious outcomes [69,70]. Furthermore, in one study, administering 100 percent of estimated caloric needs in ventilated patients with systemic inflammatory response syndrome resulted in only a transient increase (54 percent increase on the first treatment day, but lost by the third day) in the T3/rT3 ratio compared with patients receiving 40 percent of caloric needs [71]. (See "Laboratory assessment of thyroid function", section on 'Laboratory tests used to assess thyroid function'.)

SUMMARY AND RECOMMENDATIONS

●Thyroid function should not be assessed in seriously ill patients unless there is a strong suspicion of thyroid dysfunction, since there are many other factors in acutely or chronically ill euthyroid patients that influence thyroid function tests. (See 'General principles' above.)

●Many hospitalized ill patients (especially those in an intensive care unit [ICU]) have low serum concentrations of total thyroxine (T4) and triiodothyronine (T3), and their serum thyroid-stimulating hormone (TSH) concentrations are typically low, but may be low-normal or normal. (See 'Changes in thyroid hormone tests' above.)

●When thyroid dysfunction is suspected in critically ill patients, measurement of serum TSH alone is inadequate for the evaluation of thyroid function. We suggest initial measurement of TSH and free T4. Some clinicians also measure a total T3 and total T4 at the time of the initial testing to expedite decision making. However, the diagnosis may still be in doubt. (See 'Evaluation' above.)

●In critically ill patients without a strong clinical suspicion of thyroid disease and minor TSH abnormalities (eg, normal or low free T4 with TSH between 0.05 and 0.3 mU/L or TSH between the upper limit of normal and <20 mU/L), we repeat thyroid tests (TSH, free T4), typically in one to two weeks. (See 'Evaluation' above.)

●In critically ill patients with suspected hyperthyroidism (TSH usually <0.01 but can be as high as 0.05 mU/L, and normal or high-normal serum T4 and/or T3), we suggest antithyroid drug therapy (thionamides) (Grade 2C), with a plan for reassessment after recovery from the nonthyroidal illness. (See 'TSH subnormal' above and "Thionamides in the treatment of Graves' disease", section on 'Initiation of therapy'.)

●Critically ill patients with suspected hypothyroidism and TSH ≥20 mU/L with low free T4 low should be treated with thyroid hormone replacement and reassessed after recovery. In the absence of suspected myxedema coma, repletion should be cautious, beginning with approximately half the expected full replacement dose of T4 (levothyroxine). (See 'TSH high' above.)

●In critically ill patients with low free T4 and total T3 who do not appear to have an underlying primary thyroid disorder, we recommend not treating with thyroid hormone (Grade 1B). (See 'Management' above.)

●In previously euthyroid patients undergoing coronary artery bypass graft (CABG), we recommend not treating with thyroid hormone in the immediate postoperative period (Grade 1A). (See 'Management' above.)