Can We Link Thyroid Status, Energy Expenditure, and Body Composition to Management of Subclinical Thyroid Dysfunction? Free

Controversy persists over the significance of variations in TSH levels within the normal reference range and whether subclinical hypothyroidism should be treated with levothyroxine (LT4). Recent guidelines from the American Thyroid Association for the therapy of hypothyroidism (1) acknowledge that the target for TSH during treatment should be within an age-specific reference range determined from the reported data indicating cardiovascular and other benefits. Adverse cardiovascular outcomes have been especially noted with TSH levels >10 miU/L, with some effects noted when the TSH level is >7 miU/L (2). However, a recent study reported no benefit of LT4 treatment for such patients (3). An age-specific aspect has been emphasized because LT4 therapy might be associated with lower risk of ischemic heart disease but only in those aged <70 years (4), with a lower risk found for untreated subclinical hypothyroidism in patients of more advanced age (>85 years old) (5). Studies demonstrating cardiac bioenergetic impairment in subclinical hypothyroidism (6) have provided support for a substrate for cardiac dysfunction, and longstanding mild thyroid hypofunction in children might initiate proatherogenic abnormalities that lead to later cardiovascular disease (7). However, in one recent historical cohort of 1192 subjects with both subclinical hypothyroidism and heart disease, treatment with LT4 had no salutary effect on the frequency of major cardiac events or all-cause mortality (3). Furthermore, the importance of maintaining normal thyroid function has been observed in child-bearing women, illustrated by the association of infertility with greater TSH levels (8) and evidence that LT4 therapy can reduce the frequency of preterm delivery (9).

Early studies in rats have provided conclusions that could have been anticipated regarding the effects of thyroid status on energy metabolism. Rats rendered thyrotoxic with high doses of LT4 had increases in total energy expenditure by 38%, resting energy expenditure by 61%, and food intake by 18%. Animals made hypothyroid with methimazole exhibited a 10% reduction in total energy expenditure and a 12% reduction in resting energy expenditure without a change in food intake (10).

It is when variations of TSH within the reference range meet the definitions of either subclinical hypothyroidism or subclinical hyperthyroidism that a clinical issue with therapeutic implications arises. Various pathophysiological mechanisms underlie the relationship of subclinical hypothyroidism to energy homeostasis and composition. These include the direct role of TSH in brown adipose tissue and thermogenesis (11); the positive association found in obese subjects between TSH and leptin; and the potential of leptin to stimulate pituitary TRH gene expression and inhibit deiodinase type 2 activity and, thus, modify the feedback of T3 on TSH secretion (12). Also related is the downregulation of thyroid hormone receptors in obese subjects that can reduce thyroid hormone action and further increase plasma TSH, thereby constituting a form of peripheral thyroid hormone resistance (13), and the role of other thyroid hormone metabolites such as 3,5 T2 in mitochondria and cellular metabolism that influences the resting metabolic rate and weight change in both animals and humans (14, 15).

Samuels et al. (16) launched a series of important human studies in recent years in an attempt to resolve the question of the importance of variations of TSH clinically. It is these studies on which we have focused because they indirectly addressed the issue of treatment of subclinical hypothyroidism. Using methods to examine alterations in energy expenditure and body mass, they reported that a cohort of hypothyroid subjects receiving LT4 maintained at TSH levels <2.5 did not differ from those maintained at a higher TSH in regard to resting energy expenditure, body mass index (BMI), body fat mass, or visceral fat mass (16). Similarly, Garin et al. (17) reported that subclinical hypothyroidism did not alter the lean body mass, fat mass, or percentage of body fat. Samuels et al. (16) observed a direct correlation between free T3 levels and resting energy expenditure, raising the question of whether any differences in BMI, body composition, and resting energy expenditure relate primarily to free T3 and hence to deiodinative rates of the administered LT4. They noted the reported positive correlation of free T3 with BMI and body fat based on the contribution that insulinemia and bile acid can have in obese patients in stimulating D2-mediated T3 production (16, 18, 19). In contrast, al-Adsani et al. (20), in a smaller study cohort, noted that changes in the LT4 dose correlated with changes in the TSH levels and were associated with changes in resting energy expenditure. However, the increases in resting energy were not associated with alterations in the serum total T3 levels (20).

In what appears to be the same cohort of subjects reported earlier, Samuels et al. (21) examined the effect of adjustments in LT4 dosage on mood, quality of life, and cognition. The three arms of the study targeted TSH ranges of 0.34 to 2.50, 2.52 to 5.60, or 5.61 to 12.0. However, no substantial differences in the outcomes were noted, although the subjects expressed a preference for the higher dosage arm. The results were not dissimilar from those of an earlier study from the same group (22) in which central nervous system function was unaltered by suppressive doses of LT4, although the mood and health status appeared to decline. Samuels et al. (23) questioned the proposition that the upper limit of the reference range for TSH should be lowered to 2.5 miU/L (24) and examined whether subjects at the high end of the normal range experienced any reduction in health status, cognition, or mood (23). Although the subjects exhibited somewhat better decision making at the lower end of the normal TSH range, no important decrements in health status or cognition were observed. They reached the same conclusions in a study of 539 elderly men that indicated no association between TSH or free T4 on quality of life, cognition, or mood (25). This latter finding is consistent with those from cardiovascular studies (4), indicating no benefit from LT4 in subjects aged >70 years, leading the American Thyroid Association Guidelines (recommendation 6a) to suggest that older patients have a shift to the right in TSH distribution and that the rationale for therapy for this group is ambiguous (1).

In another earlier work, Samuels et al. (26) observed that LT4 replacement therapy was associated with lower resting energy expenditure, although the TSH levels remained within the reference range. They suggested that the cause could be reduced free T3 levels (26). Low serum T3 levels have been shown to correlate with hypothyroid scores on clinical assessment tools (27) and to directly influence energy expenditure during weight changes in patients with anorexia nervosa (28). The present commentary focused primarily on their most recent report (29), with studies again of the same patient cohort, in which they described the effects of an altering LT4 dosage on energy expenditure and body composition, concluding that adjustments in LT4 dosage with associated shifts in serum TSH do not have “major effects on energy expenditure or body composition.”

However, aspects of their data might suggest otherwise, such as the observed association of increases in free T4 and decreases in TSH with increases in resting energy expenditure. These findings could suggest that changes in thyroid function in and near the reference range are related to changes in resting energy expenditure. Also, because 33% of the subjects did not achieve their intended TSH target, the investigators performed an analysis based on the actual TSH levels at the end of the study. Also, they reported the lean body mass was 1 kg greater in the high-normal TSH group compared with the low-normal TSH group after adjustment for covariates (P = 0.003), which could be interpreted as based on the relationship of thyroid status to body composition. Thus, their conclusion can be called into question, even notwithstanding the well-planned structure of their study. Potential confounders to the results and conclusions might apply that relate to both controllable and uncontrollable variables, such as patient age, sex, presence of deiodinase polymorphisms, study duration, compliance, simultaneous nonthyroidal systemic illness, etiology of hypothyroidism, diet, and physical activity. Moreover, the methods for measurement of energy expenditure have generally been considered as not readily available and not well standardized.

Additional aspects of this latest study could dampen one’s enthusiasm for the significance of the somewhat marginal results. The patient cohort was heterogeneous, and the LT4 dose requirements and weight changes will vary with the cause of a patient's hypothyroidism (30–32) (i.e., surgical vs Hashimoto disease with residual endogenous functioning tissue). Athyreotic postoperative patients will require a higher LT4 dose than patients with Hashimoto thyroiditis. Also, despite being in a euthyroid state with LT4 treatment, the weight gain has been reported to be different between the two groups (32). Variability in patient compliance resulting in TSH fluctuations is an expected issue in such clinical studies and the ability to manage this by pill counts and monitoring patient diaries such as was performed in their study might not have been fully successful. All or any of these variables could have affected the investigators’ ability to establish and maintain the desired TSH level according to a given dose of LT4 for a given stable interval. Anyone who has attempted a similar clinical investigation will be aware of the often inexplicable variations in measured parameters that appear to oscillate outside of their expected or desired range. Hence, their acknowledgment that only 51%, 59%, and 60% of the subjects were in the target TSH range at the 6-, 12-, and 18-week interim visits, respectively. Resting energy expenditure can be difficult to assess accurately with variations secondary to both physical activity and diet; and the latter was not rigorously controlled in the present study. The total energy expenditure was measured using an isotopic double-labeled water technique. However, owing to supply difficulties, the measurements were performed for only 65 of the subjects. Also, the estimates of the thermic effect of food calculated from the resting energy expenditure by indirect calorimetry were performed in 80 subjects.

Thus, where do this study and, indeed, this body of work leave us regarding the entity of subclinical thyroid disease and whether LT4 treatment might have beneficial effects on metabolism, weight, mood, and cognition? We would suggest that the controversy remains largely unresolved. The investigators concluded that their results provide no support for the notion that increasing the LT4 replacement dosage might facilitate weight loss, a finding with which we have no argument. However, to extrapolate from their findings and conclude that no substantial metabolic changes will accompany alterations in free T4 and resultant TSH and the use of their observations on changes in body weight and composition as a surrogate rationale for justifying not treating patients with subclinical hypothyroidism would appear to be an unjustifiable stretch. We would reserve judgment, especially given the technical uncertainties in the cited studies and the risks attendant to untreated subclinical thyroid disease. Insofar as subclinical hyperthyroidism is concerned, it is reassuring that the group’s earlier study (26) did not indicate an adverse effect of long-term LT4 suppressive therapy on energy expenditure or body composition. However, we believe it unwise to ignore TSH levels at the extreme lower end of the reference range (i.e., subclinical hyperthyroidism) in view of the greater susceptibility of the elderly to adverse effects of excessive LT4 on the heart (33) and bones (34). At the opposite end of the age spectrum, children with subclinical thyroid disease are at risk of abnormal growth and neurocognitive development and suspiciously high or low TSH levels should not be ignored. Thus, the answer to the question posed in the title of the present commentary remains elusive, and we fully concur with the investigators’ suggestion that further long-term, carefully controlled, and comprehensive studies on selected specific populations would be of value.

Abbreviations:

Abbreviations:
 
• BMI

  body mass index

 
• LT4

  levothyroxine

Acknowledgments

Disclosure Summary: The authors have nothing to disclose.

References