In the ambulatory patient, the diagnosis of hypothyroidism is readily confirmed by finding a subnormal serum thyroxine (T4) value. However, since T4 is bound to specific serum-binding proteins, the serum T4 concentration may be altered in the absence of thyroid disease in patients who have alterations in the concentration of serum-binding proteins, particularly T4-binding globulin (Table 1). Frequently encountered clinical settings with increased serum protein binding of thyroid hormones include pregnancy and the use of oral contraceptives. Estimation of the free T4 concentration is, therefore, recommended instead of total serum T4 as the initial measurement of serum thyroid hormone concentration since the free T4 estimate will be in the normal range in euthyroid persons despite increased or decreased binding proteins. As would be expected, significant decreases in levels of serum free T4 are generally observed in hypothyroidism.
| Increase | Decrease |
| Oral contraceptives | Testosterone |
| Pregnancy | Corticosteroids |
| Estrogens | Severe illness |
| Infectious hepatitis | Cirrhosis |
| Chronic active hepatitis | Nephrotic syndrome |
| Neonatal state | Inherited |
| Acute intermittent porphyria | |
| Inherited |
A number of different methods for free T4 assays are available for estimation of free T4 concentrations. However, some of these free T4 assays perform with variable accuracy in different clinical settings. For most purposes, a free T4 estimate may be obtained by a free T4 index, an analogue free T4 assay, or any one of the other commonly used clinical chemical techniques. The free T4 index is obtained via the total T4 concentration and the thyroid hormone-binding ratio. The advantages and some pitfalls in these methods are discussed in a concurrent report (Ian D. Hay, MD, PhD, Monika F. Bayer, PhD, Michael M. Kaplan, MD, et al, unpublished data, October 1989). Throughout this report we will use the terms "free T4 estimate" and "free T4" interchangeably. Criteria for selection of free T4 assays that will be appropriate for either ambulatory or inpatient settings are being discussed by the nomenclature committee of the American Thyroid Association (Ian D. Hay, MD, PhD, Monika F. Bayer, PhD, Michael M. Kaplan, MD, et al, unpublished data, October 1989). In this regard, clinicians should require that their chemical pathologists and other clinical pathologists who direct hospital or commercial laboratories understand the complexities and limitations of different free T4 assays and employ appropriate criteria for selection of the appropriate free T4 determination.
The cause of hypothyroidism in most cases is a disorder of the thyroid gland itself. Because of the negative feedback relationship between serum thyroid hormones and pituitary thyrotropin (thyroid- stimulating hormone [TSH]) secretion, a raised serum concentration of TSH occurs in most patients with hypothyroidism. These considerations suggest that clinicians should request determinations of both the serum free T4 and serum TSH concentrations in patients who have a clinical syndrome that suggests hypothyroidism. The concomitant finding of a decrease in serum free T4 and increase in serum TSH levels confirms the diagnosis of hypothyroidism caused by thyroid gland failure; an increase in the concentration of serum TSH together with a normal or lownormal serum free T4 indicates that the patient may be at an early stage in the development of primary hypothyroidism. The etiology of thyroid failure may be established by a history of iodine 131 therapy, thyroidectomy, ingestion of antithyroid drugs, or a high frequency of thyroid disease in the family. Significant titers of antithyroid microsomal antibodies indicate that the patient suffers from Hashimoto's thyroiditis or postpartum thyroid dysfunction. In contrast to the findings in primary hypothyroidism, a subnormal serum free T4 concentration with a normal or subnormal serum TSH value suggests that the patient may have hypothyroidism secondary to a decrease in TSH secretion. This combination of results suggests a diagnosis of hypopituitarism and may warrant a detailed workup of the hypothalamic-pituitary axis by endocrinologic testing as well as radiographic procedures. Hypothyroidism caused by hypothalamicpituitary disorders occurs infrequently in comparison with hypothyroidism that results from thyroid gland failure. Moreover, the decrease in pituitary TSH secretion is usually associated with deficiencies of other pituitary hormones. The laboratory diagnosis of hypothyroidism is more complex in sick patients who are hospitalized for various illnesses. Changes in thyroid hormone secretion, distribution, and metabolism, as well as subnormal concentrations of the plasma thyroxine-binding proteins, may occur in hospitalized patients. [5] [6] Furthermore, circulating inhibitors of the binding of T4 and triiodothyronine (T3) to the serum- binding proteins have also been reported.[7] [8] These manifold and complex changes in sick individuals may result in abnormal results for various tests of thyroid function even though the patient does not have a specific thyroid disorder. In general, the more severe the illness, the greater the degree of abnormality in findings from tests of thyroid function. Indeed, endocrine consultations are frequently requested for hospitalized patients to determine whether abnormalities in results of various tests of thyroid function are consistent with nonthyroidal illness alone or whether the patient may also have associated hypothyroidism. The abnormal circulating thyroid hormone or TSH concentrations in sick patients may be adaptations to the catabolic state; many of these changes also occur during fasting, and they revert to normal when the patient recovers. [9]
A useful approach for diagnosis of hypothyroidism in sick patients is to consider patients with mild to moderate illness separately from those with severe illness, who are often in the intensive care unit. In mild to moderate illness, the combined measurement of free T4 and TSH levels should be sufficient to diagnose pre-existing hypothyroidism. In the absence of thyroid disease, the serum free T4 estimate is generally normal or raised in mild to moderate illness,[10] and serum TSH is usually in the normal range. In such patients, the criteria for diagnosis of primary hypothyroidism or hypothyroidism caused by decreased TSH secretion are similar to those described above for ambulatory individuals. Measurements of serum T3 serve no useful purpose, since serum T3 concentration may be in the normal range in 20% to 30% of patients with hypothyroidism[11] and is subnormal in approximately 70% of hospitalized patients who prove not to have any intrinsic thyroid disease.[10]
The above guidelines for diagnosis of hypothyroidism become more difficult to apply as the patient's illness becomes more severe and/or chronic. Thus, serum T4 concentration may be subnormal in more than 50% of patients in an intensive care unit, and free T4 estimates by conventional assays, such as free T4 index, are often decreased as well.[12] [13] One study has shown that a free T4 determination by equilibrium dialysis will clearly distinguish euthyroid patients with hypothyroxinemia caused by severe illness from hypothyroid patients with serum T4 levels that are decreased to a comparable degree.[14] Criteria for selection of the appropriate free T4 assay to be used in the severely ill patient have recently been discussed (Ian D. Hay MD, PhD, Monika F. Bayer, PhD, Michael M. Kaplan, MD, et al, unpublished data, October 1989).[14] Thus, in patients with critical illness, simultaneous primary hypothyroidism should be diagnosed by finding a decrease in serum free T4 and an elevation in serum TSH levels. A potential pitfall in this approach is the frequent use of dopamine and pharmacologic doses of corticosteroids in critically ill patient populations. These agents, commonly used in the intensive care unit, may suppress an elevated serum TSH concentration toward or into the normal range.[15] [16] Nevertheless, only a rare case of primary hypothyroidism does not exhibit significant elevation in serum TSH concentration.
Another possible pitfall in the diagnosis of primary hypothyroidism in the critically ill patient is the transiently raised concentration of serum TSH that is found in many euthyroid patients as they recover from severe illness.[9] [17] Values of serum TSH that are in the range previously considered diagnostic of primary hypothyroidism may occur in this setting; TSH generally returns to the normal range within a few days or weeks. Thus, the finding of low-normal or slightly decreased serum free T4 concentrations in conjunction with increased concentrations of serum TSH when using conventional free T4 assays is consistent with a diagnosis of either primary hypothyroidism or the euthyroid state in a patient who is recovering from critical illness. Since unrecognized primary hypothyroidism occurs in only about 1% of the adult population, most patients with these findings will be euthyroid. Clinical factors such as previous treatment for Graves' disease, thyroid surgery, significant titers of antithyroid microsomal antibodies, or findings on physical examination typical of hypothyroidism all would suggest that the elevation in serum TSH concentration is secondary to thyroid failure.
Only a rare patient with critical illness may have hypothyroidism that results from hypothalamic or pituitary disease associated with decreased TSH secretion; such patients will often have signs of hypogonadism and hypoadrenalism as well. The pituitaryadrenal axis should be evaluated in such patients by measuring serum cortisol, since an undiagnosed deficiency in cortisol could have grave clinical consequences in sick patients. The secondary hypothyroidism in these critically ill patients will be established by finding serum free T4 in the hypothyroid range and serum TSH in the low-normal range or decreased. The apparently paradoxical finding of measurable serum "sensitive" TSH in some cases of secondary hypothyroidism apparently results from secretion of TSH that has reduced biologic activity. In the absence of hypothalamic or pituitary disease, a normal TSH value excludes primary thyroid disease, provided that the patient is not receiving dopamine.[16] The sensitive TSH assay is a new method to measure TSH and provides much greater sensitivity than the radioimmunoassay method (Ian D. Hay, MD, PhD, Monika F. Bayer, PhD, Michael M. Kaplan, MD, et al, unpublished data, October 1989).