Hi, my name is Dr. Kari Sims and I am the current Ogle Jonason Professor of Surgery and the Chief of Trauma Critical Care and Burn at The Ohio State University. Today I'm going to talk about thyroid dysfunction and critical illness and whether we should treat or not treat. The objectives of this talk are to discuss the incidence of non-thyroidal illness syndrome in critically ill patients, to review the pathophysiology of its acute and chronic forms, to review the literature supporting or refuting thyroidal supplementation, and finally to give some recommendations based on the literature. Thyroid function impacts every physiologic system and hypothyroidism can mimic critical illness in many ways including decreased mental status, cardiac dysfunction, hypothermia, and impaired immunologic function. It is estimated that over 80% of patients presenting to the ICU will demonstrate some evidence of abnormal thyroid function either on admission or during their hospital stay in the intensive care unit. This talk will not discuss treating hypothyroidism that is previously known, myxedema coma, or cyrotoxicosis. Instead we will talk about non-thyroidal illness syndrome as it tends to be the most controversial. Non-thyroidal illness syndrome or otherwise known as euthyroid 6 syndrome is characterized by normal or decreased TSH as well as concomitant decrease in free T4 and free T3 levels. Non-thyroidal illness syndrome is observed in approximately 70% of ICU patients and it is regardless of their COVID status. So is non-thyroidal illness syndrome helpful or harmful? In order to answer this question we must first look at the pathophysiology of acute non-thyroidal illness syndrome. There is a critical illness and the response is primarily tissue based. Acutely there is decreased binding proteins in a response of the acute phase response and there is also decreased peripheral hormone uptake with decreased expression of thyroid hormone receptors as well as decreased importing mechanisms. There is also alterations in the expression and activity of key deionases in the periphery with suppression of D1, the primary enzyme responsible for converting the inactive T4 to the active T3 as well as an increase in D3 which converts D4 to its inactive form of reverse T3 as well as converting T3 into T2, another inactive form. This leads to an abrupt decrease in energy expenditure which might be helpful for the organism as well as in granulocytes it increases the available iodine necessary for increased bactericidal activity. Interestingly a similar pattern is seen when healthy volunteers fast, they pretty much look like they have acute non-thyroidal illness syndrome and it begs the question that are the findings that we see in critically ill patients early on in their stay related really to decreased thyroid function or is this an adaptive response to decreased caloric intake? The adaptation however goes away. So here in a very interesting study done by Grete Vandenberg they had healthy volunteers as well as acutely ill and chronically ill patients monitored overnight from 9 o'clock at night to 6 o'clock in the morning and they took serial blood samples to look at the expression of TSH in the serum. What they found was that healthy volunteers have nocturnal variation with peaks and valleys of TSH. In acute critical illness however although the TSH levels are appropriately elevated you lose that variability and finally with prolonged critical illness defined as anything greater than 10 days of ICU stay you see that there is a significant decrease in the expression of TSH as well as complete loss of diurnal variation. This leads us to chronic non-thyroidal illness syndrome and its pathophysiology. This is primarily a hypothalamic response and a neuroendocrine failure. Thyrotropin-releasing hormone expression is suppressed in prolonged critical illness although we don't understand the mechanism why. We do know that when patients are examined post-mortem those that have died of critical illness have decreased mRNA expressions of thyrotropin-releasing hormone whereas those that have died of acute insults have normal expression. So there is a clear difference in the mechanism leading to decreased TRH expression. This gives us a hormonal pattern that is quite similar to central hypothyroidism meaning that TRH is decreased, the TSH also decreased. This leads to decreased T4 expression from the thyroid gland as well as tissue effects. Interestingly there is some data to suggest that there is actually increased transporter activity in chronic non-thyroidal illness as well as increased D2 deiodinase expression in the tissues. However there is interesting data to suggest that D2 polymorphisms have been linked to decreased susceptibility of sepsis in patients. So again is non-thyroidal illness syndrome helpful or harmful and does it depend on where we are in its temporal expression? In general we think of acute non-thyroidal illness syndrome as potentially protective and beneficial. However in trauma patients this may not be the case. In a prospective observational study of 18 hypotensive severely injured patients presenting in hemorrhagic shock, serial blood samples were taken first in the trauma bay then in the ICU after hemorrhage control and then 24 hours and 48 hours later. What these authors found was that lower free T4 levels on arrival actually predicted an increased need for blood product utilization and crystalloid use. And in fact there was a trend toward higher injury severity score and vasopressor use in patients who had lower free T4 levels. Similarly lower free T3 levels on arrival were associated with increased blood product use utilization. However over the 48 hours all patients had a dramatic decrease in their T3 levels reflecting severe injury. This relationship between thyroid function and the ICU also correlates with mortality in all comers. In this study done by Dr. Wong et al, 480 consecutive patients were admitted to an intensive care unit. Thyroid functions were drawn but they were not standardized when they were taken. were found to have significantly lower T4, free T4, T3, and free T3 values. And in fact of the thyroid function levels free T3 was the best independent predictor. In fact it actually was better than CRP in this population. More recently a large meta-analysis has confirmed these findings suggesting that non-thyroidal illness syndrome predicts outcomes in critically ill patients. In this meta-analysis of 25 studies there were almost 7,000 patients. They had different time points of measurement however the median prevalence of non-thyroidal illness syndrome was 58% in all comers. These studies found that both T3 and free T3 were significantly lower in non-survivors and they also found that free T4 was significantly lower in non-survivors when compared to survivors. They conclude that non-thyroidal illness syndrome is associated with increased risk of death with an odds ratio of 2.2. In this diagram you can see a number of the studies and the odds ratio in them range from 1.86 to 12.43 suggesting that non-thyroidal illness syndrome really does predict increased mortality. So given this increased mortality should we treat or should we not treat? Unfortunately for the last 40 years the recommendation has been to not treat and this is primarily based off of a very small study of 23 MICU patients who presented with T4 levels that were less than 5. In this small study 11 were randomized to the treatment arm and 12 were maintained as controls. Those in the treatment arm received intravenous T4 for two weeks versus placebo. Not surprisingly the T4 and free T4 levels normalized by day 5. However the TSH levels also decreased. Not surprisingly in the small group there were no differences in survival. Thus the authors conclude that there was no benefit and potential harm because the TSH was decreased. There are several potential issues with this study however. There certainly wasn't power to detect differences in mortality and I don't think it would actually be published today and in fact maybe T4 may not be the right treatment especially since there is the concern of differential deiodinases in their expression. Since then there has been a variable potpourri of clinical studies. In fact there's been more than 20 randomized control trials looking at treating non-thyroidal illness syndrome in a variety of different conditions. These studies look at different ages, anything from neonates to older folks undergoing coronary artery bypass. They look at different disease processes, everything from prematurity to burns to critical illness. They look at different treatments whether it's T3 or T4 or even thyrotobone stimulating hormone and they have different durations and different endpoints. So clearly it is a whole hodgepodge of clinical trials and that we cannot actually discern anything because they are all underpowered to look for differences. Thus there is no conclusive benefit to treating non-thyroidal illness syndrome. That being said there's also no harm and in fact some have argued that given the high mortality associated with non-thyroidal illness syndrome, what's the harm in treating? I would like to talk about one therapeutic modality in particular and this really tackles the issue of the neuroendocrine failure at the hypothalamic level. Looking back at Dr. Vandenberg's trial you can see here again the tracing of the normal TSH during night time with its peaks and its valleys, the acute care illness tracing of TSH which is elevated but flat and then the prolonged critical illness which is depressed without any variation. In order to address this she looked at giving critically ill patients who had been in the ICU for 15 to 18 days a short duration of thyrotropin releasing hormone. They started this at night time and it was a continuous infusion of one microgram per kilogram per hour and what she found was that by giving patients TRH they could easily increase the levels of the TSH in the serum. However it didn't have its variability. When she added the growth hormone secretagogue, a peptide GHRP2, you can see that this had led to the return of the pulsatility and the secretion of TSH. When she looked at the downstream products, that of the T4, T3 and reverse T3, she noticed something incredibly surprising and here you can see the placebo whereas over the 24 hours there was a decrease in free T4, a decrease in free T3 as well as an increase in the RT3 with just the placebo. When she gave the TRH you can see that this led to a marked increase in T4, a marked increase in the T3, however there was also an increase in the reverse T3. When she gave only the growth hormone secretagogue, this pretty much looked like the placebo settings. And finally when she gave the TRH plus the growth hormone secretagogue, you can see that the T4 was increased, the T3 was increased and unlike previously, the reverse T3 actually decreased. So this normalized the whole thyroid profile. So what are the recommendations? Well, I would recommend that you continue treating hypothyroid patients who are admitted to your ICU. So if they are hypothyroid to begin with and they come in on levothyroxine, you should continue it. If you suspect that the patient might have non-thyroidal illness syndrome but you check and your TSH is actually elevated two weeks after their ICU admission, this is not simple non-thyroidal illness syndrome. In fact, it is likely that it is subclinical hypothyroidism and should be treated as such. The data regarding the treatment of non-thyroidal illness syndrome is poor at best and warrants a large clinical trial that has yet to be done. As such, there are no current recommendations for when or how we should treat non-thyroidal illness syndrome. That being said, it is associated with a high risk of mortality and if you do choose to treat, it appears to be well tolerated. Thank you very much.

thyroid dysfunction

critical illness

ICU patients

non-thyroidal illness syndrome

abnormal thyroid function