Higher Caloric Exposure in Critically Ill Patients Transiently Accelerates Thyroid Hormone Activation

Liam McKeever; Sarah J. Peterson; Omar Lateef; Sally Freels; Tatiana L. Fonseca; Barbara M. L. C. Bocco; Gustavo W. Fernandes; Kelly Roehl; Kristen Nowak; Marisa Mozer; Antonio C. Bianco; Carol A. Braunschweig


J Clin Endocrinol Metab. 2020;105(2) 

In This Article


The primary findings of this study were that higher caloric exposure in critically ill patients induced a rapid but transient increase in the plasma T3/rT3 ratio when compared with lower caloric exposure (Table 5, Figure 3a). This elevation in the plasma T3/rT3 ratio was mostly driven by an early transient increase in T3 in the higherfed group (Table 6, Figure 3b), which most likely resulted from an acceleration in extrathyroidal T3 production. Notably, this was partially reversed after 24 hours of continued higher caloric exposure with the serum T3/rT3 ratio progressively returning to what is observed in the patients kept on a lower caloric exposure.

How might these nutrition-induced changes in thyroid hormone economy be explained? The absence of significant changes in plasma T4 levels suggest that the alterations in thyroid economy took place predominantly outside the thyroid gland. Such changes are likely to include an acceleration in extrathyroidal T3 production via the deiodinase 1 (D1) and/or deiodinase 2 (D2) pathways, which would also explain the reduction in plasma rT3 levels. Despite the transient elevation in TSH in the 100% ECN vs the 40% ECN group (Table 9, Figure 4), it is less likely that thyroidal T3 production was substantially increased given that plasma T4 and rT3 levels remained unaffected. Indeed, TRH infusion to NTIS patients has been shown to elevate circulating T3 levels, but T4 and rT3 levels were elevated as well.[32]

It is well known that the D1 pathway, particularly in the liver, is stimulated by caloric exposure.[20,33] In animal models of caloric restriction or starvation and in critically ill patients, the commonly observed drop in circulating T3 levels is associated with a reduction in hepatic D1 activity.[9] This can be explained due to the sensitivity of liver D1 to dietary carbohydrates. However, liver D1 is also positively regulated by circulating T3,[34] making it difficult to ascertain how much of the elevation in D1 activity is a cause or a consequence of the elevation in circulating T3. Thus, it is conceivable that the higher caloric exposure in the 100% ECN group activated the hepatic D1 pathway, which would also explain the drop in plasma rT3 observed on study day 1 (Table 7, Figure 3C). Unfortunately, measuring hepatic D1 activity requires a liver biopsy, beyond the scope of the present investigation. At the same time, we know that the D2 pathway, the main mechanism for extrathyroidal T3 production in humans,[35] is also activated by caloric exposure.[36] D2 is expressed at high levels in the brain in the pituitary gland, none of which is thought to contribute significantly to the pool of circulating T3. However, D2 is expressed at low levels in large organs and tissues such as skin, skeletal muscle, and endothelium, which collectively contribute to the circulating pool of T3.[7] In these tissues, DIO2 expression is normally kept to a minimum due to FOXO1 binding to the DIO2 promoter, slowing down conversion of T4 to T3. Feeding stimulates insulin secretion and activates the PI3K-mTORC2-Akt pathway, which excludes FOXO1 from the cell nucleus and de-represses DIO2, accelerating D2 synthesis and conversion of T4 to T3.[36] This is a likely explanation for the elevation of plasma T3 in the 100% ECN group of patients.

Prospective exploration regarding the relationship between nutrition and thyroid parameters in critically ill patients is sparse and challenging to interpret. To date, only 1 PRCT (n = 59) assessing NTIS in critically ill patients exists.[24] This study explored the effects of early (initiated within 24–48 hours of ICU admission) vs delayed (withheld for at least 48 hours) EN in 59 patients with traumatic brain injury. The effect of real-time calorie exposure in this study is difficult to interpret, as the calorie exposure data for the 24 to 48 hours preceding blood sample acquisition was not reported. Blood sample acquisition occurred on days 6 and 12. Any effects were inferred to be caused by differences in calorie exposure that occurred days prior on days 1, 3, 7, and 10 of the study. With this in mind, their early enteral group had higher levels of free T3, T4, and TSH on study day 6 vs 0 compared with the patients in the delayed enteral group. The only other study to explore this in critically ill patients was a subanalysis[25] (n = 280) on the multicenter EPaNIC study[37] (n = 4640) which explored the effect of supplemental PN vs late PN on mortality and complication rates in ICU patients. Patients in the early PN group received a 20% glucose solution with a goal of receiving 400 kcal on study day 1 and 800 kcal on study day 2. The subanalysis was a sampling of patients for whom EN was surgically contraindicated. Blood was obtained on study day 3, however it is unknown if these specimens were taken prior to initiation of full PN. By design, the late PN group was unable to receive EN, and thus were essentially fasted. They found increased T3/rT3 ratio, TSH, T3, T4, and a decreased rT3 in patients that received supplemental early PN compared to those in the delayed PN group, further supporting the hypothesis of an attenuated NTIS in critically ill patients with higher nutrition exposure.

While starvation-induced alteration in thyroid hormone economy is thought to be a positive adaptive mechanism of energy conservation, the consequences of NTIS in critical illness are unclear.[10] The current study is unable to elucidate whether attenuating NTIS in these patients was beneficial or harmful, or neither, but it demonstrates that feeding has real-time causal consequences to the body's physiologic response to illness. To our knowledge, this has only been addressed in an exploratory fashion through the EPaNIC subanalyis, in which the authors qualitatively demonstrated through a preliminary mediation analysis that the changes in the T3/rT3 ratio mediated the beneficial effect of delayed nutrition exposure on the likelihood of early alive discharge from the ICU. Depression of the T3/rT3 ratio further improved the benefit in this model.[25] A formal mediation analysis would be necessary to determine if indeed this mediation was statistically significant, as discussed in.[38] PRCTs designed and powered to specifically explore the effect of this relationship on clinically relevant outcomes are urgently needed.

Strengths and Limitations

This study has several strengths. First, it is a prospective, randomized controlled feeding trial in critically ill patients that included assessment of plasma levels of rT3 to enable analysis of thyroid hormone metabolism. Second, its repeated measures design provided the power and detail necessary to fully explore the effect of feeding on NTIS over multiple timepoints. Third, patient groups received the same type of nutrition exposure, varying only in the level of kcal/kg received. Fourth, by using SIRS as inclusion criteria, the interference caused by inflammatory status was minimized, enabling a focus on the effects of nutrition on NTIS. Limitations associated with this trial include its relatively small sample size, which limits generalizability. Additionally, while patients in the 100% ECN group received significantly more calories on average, high levels of feeding were difficult to achieve in some patients, leading to increased homogeneity in calorie exposure between the groups, especially on days 4 and 5. This did not likely explain the attenuation in our T3/rT3 ratio, however, as models restricted to the first 3 study days demonstrated the same significant rise and attenuation in T3/rT3 ratio. This also makes it unlikely our results were an artifact of missing data from patients who died or were discharged prior to completing 7 days in the study. Finally, while both the patients and the lab technicians were blinded to group allocation status, the medical staff was not.

In conclusion, the present study demonstrated that higher calorie exposure caused an acute attenuation in NTIS in critically ill SIRS patients, characterized by an elevation in the plasma T3/rT3 ratio, which reversed itself over 3 days. Exogenous nutrition produces immediate changes in our physiologic response to critical illness. The consequences of these alterations are largely unexplored and warrant further investigation. Future research should explore this in a population large enough to examine its association with clinically relevant outcomes such as mortality, nosocomial infection, time on mechanical ventilation, and length of stay.