ENDOCRINE PRACTICE Rapid Electronic Article in Press Rapid Electronic Articles in Press are preprinted manuscripts that have

been reviewed and accepted for publication, but have yet to be edited, typeset and finalized. This version of the manuscript will be replaced with the final, published version after it has been published in the print edition of the journal. The final, published version may differ from this proof. DOI:10.4158/EP13460.OR 2013 AACE.

Original Article

EP13460.OR A DIAGNOSTIC SCORING SYSTEM FOR MYXEDEMA COMA

Running title: Diagnostic Scoring for Myxedema Coma Geanina Popoveniuc, MD1, 2, Tanu Chandra, MD3, 4, Anchal Sud, MD1, Meeta Sharma, MD1, Marc R. Blackman, MD 2, 4, 5, Kenneth D. Burman, MD1, Mihriye Mete, PhD 6,7 , Sameer Desale, MS 6,7 , Leonard Wartofsky, MD1

From: 1Division of Endocrinology, Department of Medicine, MedStar Washington Hospital Center, Washington DC; 2Division of Endocrinology, Department of Medicine, Georgetown University Hospital, Washington DC; 3Division of Endocrinology, Department of Medicine, Veterans Affairs Medical Center, Washington DC; 4Division of Endocrinology, Department of Medicine, George Washington University Hospital, Washington, DC; 5Research Service (151), Veterans Affairs Medical Center, Washington DC; 6 Department of Biostatistics and Bioinformatics, Medstar Health Research Institute, Hyattsville, MD; 7 Georgetown-Howard Universities Center for Clinical and Translational Sciences, Washington, DC (GHUCCTSCTSA)

Correspondence address: Geanina Popoveniuc MD, address: 110 Irving Street NW, 2A72, Washington, DC, 20010-2975. Email: geanynar@yahoo.com

DOI:10.4158/EP13460.OR 2013 AACE.

Keywords: myxedema coma; hypothyroidism; diagnosis; scoring system.

Abstract

Objective: To develop diagnostic criteria for myxedema coma (MC), a decompensated state of extreme hypothyroidism with a high mortality rate if untreated, in order to facilitate its early recognition and treatment.

Methods: The frequencies of characteristics associated with MC were assessed retrospectively in patients from our institutions, in order to derive a semiquantitative diagnostic point scale that was further applied on selected patients from literature. Logistic regression analysis was used to test the predictive power of the score. Receiver operating characteristic (ROC) curve analysis was performed to test the discriminative power of the score.

Results: Of the 21 patients, 7 were re-classified as not having MC (non-MC), and they were used as controls. The scoring system included a composite of alterations of thermoregulatory, central nervous, cardiovascular, gastrointestinal, and metabolic systems, and presence or absence of a precipitating event. All our 14 MC patients had a score of 60, whereas 6/7 nonMC patients had scores of 25-50. Sixteen of 22 MC patients from literature had a score 60, and 6/22 scored between 45 - 55. The odds ratio per each score unit increase as a continuum was 1.09 (95% CI, 1.01-1.16; p =0.019); a score of 60 identified coma with an odds ratio of
DOI:10.4158/EP13460.OR 2013 AACE.

1.22.The area under the ROC curve was 0.88 (95% CI, 0.65-1.00), and the score of 60 had 100% sensitivity, and 85.71% specificity.

Conclusions: The scoring system proposed indicates a score of 60 potentially diagnosing MC, whereas scores between 45-59 could classify patients at risk for MC.

Abbreviations: MC = myxedema coma; ROC = receiver operating characteristic; TSH = thyroid stimulating hormone; T4 = thyroxine; T3 = triiodothyronine; GCS = Glasgow Coma Scale APACHE II = Acute Physiology and Chronic Health Evaluation; SOFA = Sequential Organ Failure Assessment; SD = standard deviation; MWHC = Medstar Washington Hospital Center; VAMC = Veterans Affair Medical Center.

Introduction Myxedema coma is a rare form of extreme hypothyroidism with a mortality rate that may approach 60% [1]. The condition represents a state of decompensated hypothyroidism that usually occurs after a period of longstanding, unrecognized or poorly controlled thyroid hypofunction and is often precipitated by a superimposed systemic illness. Such precipitating or exacerbating factors include infection, trauma, certain medications, hypothermia, cerebrovascular accident, congestive heart failure, metabolic disturbances, and electrolyte abnormalities [1-3]. If left untreated, the clinical course is one of multi-organ dysfunction with characteristic lethargy progressing to altered sensorium (stupor, delirium, and coma). Hypothermia is a key early manifestation in most patients and may be quite profound (less than 26 C). Respiratory depression leading to hypoventilation and hypercapnia may
DOI:10.4158/EP13460.OR 2013 AACE.

necessitate intubation and mechanical ventilation. Decreased cardiac contractility, bradycardia, cardiomegaly, and arrhythmias may lead to hypoperfusion and cardiogenic shock. Other common abnormalities seen in patients with myxedema coma include gastrointestinal dysfunction, renal impairment, hyponatremia, hypoglycemia, hypoxemia and anemia [1].

The diagnosis of myxedema coma is usually based on clinical manifestations, a history of moderate to severe hypothyroidism, and is confirmed by laboratory testing, with elevated serum thyrotropin (TSH), and decreased total and free thyroxine (T4), and triiodothyronine (T3). Early diagnosis, supportive care, and treatment with intravenous thyroxine have been shown to improve outcomes [4]. Recent reports including prospective studies [2, 3, 5] have focused on establishing predictors of poor outcome in patients with myxedema coma.

Coma on admission, lower GCS (Glasgow Coma Scale) score and an APACHE II (Acute Physiology and Chronic Health Evaluation) score of < 20 were demonstrated to be reliable predictors of higher mortality in the prospective study of Rodriquez et al. [2] of 11 patients with myxedema coma. They also noted that the mean age of survivors was lower than that of non-survivors, albeit not statistically significantly. Heart rate, body temperature, mean free T4, and mean TSH did not differ between survivors and non-survivors. Dutta et al [3], in a report of 23 patients with myxedema coma, found hypotension and bradycardia on admission, need for mechanical ventilation, hypothermia unresponsive to treatment, sepsis, intake of sedative drugs, lower GCS score, and high APACHE II and SOFA (Sequential Organ Failure Assessment) scores highly predictive of a poor outcome. Results from a Medline search of 82
DOI:10.4158/EP13460.OR 2013 AACE.

cases of myxedema coma [5] revealed that older age, cardiac complications, such as hypotension and sinus bradycardia with low voltage QRS, and high dose thyroid hormone replacement during treatment for myxedema coma were associated with a fatal outcome after 1 month of therapy. There was no significant difference in mortality based upon the APACHE II score and the presence of pulmonary complications.

The diagnosis of myxedema coma is mainly clinical, with no clear cut criteria that might distinguish either hypothyroidism alone or coma of other etiologies from true myxedema coma. In view of the high morbidity and mortality of myxedema coma [2], the development and application of criteria for its identification could allow earlier diagnosis and treatment that may have a salutary effect on prognosis for recovery and outcome. [4]

Materials and Methods

Study population Our study population was based on all patients age 18 years and older who presented to MedStar Washington Hospital Center (MWHC), Washington DC and Veterans Affair (VA) Medical Center, Washington DC from 1989 to 2009, with an admitting or discharge diagnosis of myxedema coma. Definitions The following definitions and grading systems were employed: hypothermia was defined as a temperature lower than 35C. Bradycardia was defined as heart rate less or equal to 60 beats per minute and hypotension as blood pressure less than 90/60 mmHg, or a mean arterial
DOI:10.4158/EP13460.OR 2013 AACE.

pressure less than 70. Neurological findings were graded based on the severity of mental status changes, from somnolence to obtundation, stupor and coma. Obtundation was defined as less than full mental capacity, but still easy arousable with persistence of alertness for brief periods of time [1]. Stupor was applied to the state of lack of critical cognitive function and level of consciousness, responsiveness only to painful stimuli, while coma was considered to be the state of complete lack of responsiveness. Hypoglycemia was defined as a blood glucose level < 60 mg/dL and hyponatremia was classified as a serum sodium < 135 mEq/L . To define hypoxemia we used a threshold for oxygen saturation at room temperature of less than 88% or pO2 less than 55 mmHg, while hypercapnia was indicated by a pCO2 level of 50 mmHg or greater. The diagnosis of primary hypothyroidism was based on levels of total or free thyroxine (T4) below the reference range together with an elevated serum TSH. Reference ranges were as follows: total T3 71-180 ng/dL, total T4 4.5 -12 ug/dL, free T4 0.8 1.7 ng/dL, and TSH 0.45 4.5 mIU/L. Methodology Each chart was retrospectively reviewed (by GP and TC) to note patient demographics and the clinical manifestations of myxedema coma in each patient on presentation. The following characteristics were recorded for each patient: demographics (gender, age, race, past medical history, to include history of hypothyroidism, or thyroid surgery, medications, medication non-compliance), vital signs at the time of MC diagnosis (temperature, heart rate, respiratory rate, blood pressure, oxygen saturation), respiratory status (supplemental oxygen, mechanical ventilation), neurological status (somnolence, lethargy, obtundation, stupor, coma, seizures), gastrointestinal manifestations (anorexia, abdominal pain, constipation, decreased/absent intestinal motility), laboratory findings (complete metabolic panel, TSH, free T4 and total T3,
DOI:10.4158/EP13460.OR 2013 AACE.

blood cultures, urine cultures), electrocardiographic findings, chest X Ray reports, and history of precipitating insults, if present. The frequency of various factors distinguishing myxedema coma from hypothyroidism without coma or non-thyroidal causes of coma was assessed and weighted to further develop a diagnostic point scale in order to enable a semiquantitative distinction between uncomplicated hypothyroidism, severe hypothyroidism and myxedema coma. The potential utility of the diagnostic scoring system was assessed by application to selected patients reported in the literature. Statistical analysis Microsoft excel spreadsheet software was used to note the frequency of clinical events. Baseline characteristics between the two groups (MC vs. non-MC) were compared by using Fishers exact test for categorical variables and two sample t-test for continuous variables. A p- value of <0.05 was considered to be statistically significant. Logistic regression analysis was used to test the predictive power of the score for myxedema coma. Results were expressed using odds ratio and 95% confidence interval. Further, receiver operating characteristic (ROC) curve analysis was performed to test the discriminative power of the score. The discriminative power was measured by using area under ROC curve. Sensitivities and specificities were calculated for all values of the score and the cutoff point was identified with left topmost point on ROC curve (representing the highest sensitivity and specificity). Statistical analyses were performed in SAS 9.3, SAS Institute Inc., Cary, NC, USA. The study protocol was approved by the Institutional Review Boards of MWHC and VAMC.

Results
DOI:10.4158/EP13460.OR 2013 AACE.

Chart review identified twenty one patients who had been diagnosed with myxedema coma by an endocrinologist. We re-classified seven patients as non-myxedema coma (non-MC) as we believed they were misdiagnosed with myxedema coma, and we used them as a control group (Table 3). Reasons for re-classification included normal free T4 levels and only marginally elevated serum TSH (patients 1, 2, 4, 7), or absence of any degree of mental status changes (patients 3, 5 and 6), since mental status alteration was a criteria historically used to diagnose myxedema coma in patients with hypothyroidism. The frequency of demographics and clinical characteristics of the patients in each group is presented in Table 1 and a summary of the patients clinical characteristics is detailed in Tables 2 and 3 (page 1 and 2). As noted in Table 1, there were no statistical significant differences between the two groups in terms of patient clinical characteristics, to distinguish patients with myxedema coma, from those with other forms of hypothyroidism. The age (mean SD) at presentation was 68 15 years in MC group vs. 66 23 years in non-MC group (p = 0.81), with 57% of men in MC group vs. 43% in non-MC group (p = 0.66). The distribution of the neurological alterations in MC group was relatively similar throughout the entire spectrum of neurocognitive dysfunction, with 36% of the patients described as somnolent or lethargic, and with coma being present in 29% of the subjects (Table 1).The most common clinical manifestations in MC patients were hypothermia (50% in MC vs. 29% in non-MC, p = 0.64) and hypotension (50% in MC vs. 14% in non-MC, p = 0.17). A wide spectrum of EKG alterations was noted in patients with MC, with bradycardia present in 36% of the cases. Myxedema coma patients had more frequent and wider distribution of EKG alterations, metabolic disturbances and gastrointestinal manifestations, than non-MC patients, although
DOI:10.4158/EP13460.OR 2013 AACE.

none reaching statistical significance (Table 1). Each patient was noted to have had one or more identifiable precipitating events.

Based on the above findings, we constructed a diagnostic scoring system to enable a semiquantitative distinction between uncomplicated hypothyroidism, severe hypothyroidism and myxedema coma (Table 4). The lack of statistically significant difference between all the clinical characteristics of the two groups, combined with the wide and relatively similar distribution of events in each category led to the construction of a comprehensive multisystemic diagnostic scale, in which points were assigned using a stratified approach based on the severity of each condition in a particular system. The highest weighted description applicable in each category was considered and scores were totalled. When a given descriptive characteristic was encountered in more than one category (i.e., precipitating event and metabolic disturbance), the condition was counted once.

When applied to the fourteen patients with MC, a score of 60 or higher (60 - 120) was calculated to be diagnostic of myxedema coma (Table 2, page 2). Six of the seven patients with non-MC had scores ranging between 25 and 50 (Table 3, page 2). A single patient from this latter cohort had a score of 110, but he was excluded because of a normal free T4 of 1.14 ng/dL and an only mild TSH elevation.

Logistic regression univariate analysis identified the score as a continuum to be predictive of the outcome with an odds ratio of 1.09 per unit of the score (95% CI, 1.01-1.16; p =0.019). A score of 45 predicted coma with a probability of 0.27 and an odds ratio of 0.37, respectively,
DOI:10.4158/EP13460.OR 2013 AACE.

whereas a score of 60 had a predictive probability of 0.55, with an odds ratio of 1.22. The model overall was significant (Chi-square test p-value = 0.0006). The area under the ROC curve of the prediction score was 0.88 (95% CI, 0.65 1.00) (Fig 1). The cutoff point on ROC curve corresponded to the score of 60, which had the highest sensitivity (100%) and specificity (85.71%), with a positive likelihood ratio of 7.0 and negative likelihood ratio of 0.0. The score of 45 had 100% sensitivity, but a lower specificity of 42.86%, whereas a score of less than 25 had 0% specificity (Fig 1).

When applied to patients in the literature for whom enough clinical data were available, the diagnostic scoring system identified 16 out of 22 patients as having myxedema coma (score 60) (Table 5). The remaining six patients would have been classified as being at risk for myxedema coma (scores ranged between 45 - 55), but did not quite meet the criteria for a diagnosis of myxedema coma. None of the twenty two patients had scores at presentation that qualified them as unlikely to have myxedema coma.

Discussion

Although it is generally accepted that the diagnosis of myxedema coma should rely on some degree of mental status alteration, impaired thermoregulatory response and the presence of a precipitating event [6], clear cut diagnostic criteria to define myxedema coma have not been established. Moreover, uncertainty of diagnosis is suggested by the numerous hypothyroid patients with presumed myxedema coma reported in the literature in whom at least one of these features was minor or absent.
DOI:10.4158/EP13460.OR 2013 AACE.

Although altered mental status was a prominent aspect of the presenting clinical picture in all our patients, it would be tenuous to base a diagnosis on this alone. There may be innumerable etiologies for mental status change, but it is through combination with other signs and symptoms of our scoring system, along with thyroid function test results, that the mental status changes allow a more precise focus on the diagnosis of myxedema coma. To our knowledge, there have been no previous reports of clinical algorithms to define diagnostic criteria for myxedema coma, likely due to the paucity of cases and consequent lack of studies to address this issue. Accordingly, we have developed a diagnostic scoring system for myxedema coma, and assessed its potential utility in a cohort of patients from our two institutions, as well as applying it to selected patients identified in the literature [2, 1323]. Our hope is that this scoring system will enable earlier diagnosis and treatment of patients with myxedema coma. Importantly, most of the patients whom we evaluated from the literature were likely underscored due to limited clinical data availability. Thus, an assigned score of 60 could easily have been achieved with one or two more variables being present, such as the lacking details of metabolic abnormalities, EKG changes, and/or gastrointestinal manifestations. Patient 14 [Table 5] [15] was of particular interest, as she initially presented to the hospital with biochemical evidence of subclinical hypothyroidism, and clinical features that would not have diagnosed her with myxedema coma, given a score of 40. Shortly after admission, her clinical status deteriorated and she was diagnosed with myxedema coma, achieving a score of 80, based on our diagnostic scale. Of note, the patients biochemical markers continued to reflect a state of subclinical hypothyroidism throughout her hospitalization, showing that a reliance on thyroid function tests alone could have potentially missed the development of
DOI:10.4158/EP13460.OR 2013 AACE.

myxedema coma, thereby delaying diagnosis and treatment of this patient.

The predictive power of the score as a continuum showed an odds ratio of 1.09 (95% CI, 1.011.16; p =0.019) suggesting that with each unit increase in the score within the range of available data, the odds of myxedema coma increases by a factor of 1.09, or by 9%. For instance, a change in score from 50 to 51 would change the predictive probability of coma from 0.35 to 0.37, or from odds ratio of 0.54 to odds ratio of 0.58. The score of 60 represented a turning point and predicted coma with a high accuracy, given its predicted probability of 0.55, which conferred an odds ratio of 1.22. The odds of coma for a score of 45 was approximately 1/3 (0.37), which corresponded to a predicted probability of 0.27. The discriminative power of the scoring system was high, with area under the ROC curve of 0.88 (95% CI, 0.65 1.00). The score of 60 had the highest sensitivity (100%), and specificity (85.71%) of the scores calculated which makes it a good screening tool given the highest sensitivity and the relatively high specificity. The score of 45 had 100% sensitivity, but a lower specificity of 42.86%. Given the above considerations, we propose that with application of the recommended scoring system, a score of 60 or higher will be highly suggestive of myxedema coma, a score between 45 and 59 will represent risk for myxedema coma, and that a score of less than 45 is unlikely to indicate myxedema coma. Given the small sample size, our model was not capable of producing a threshold score for patients at risk for myxedema coma, therefore the scores between 45-59 are only our suggestion of representing patients in this category, based on the given probabilities.

Neurocognitive dysfunction in patients with myxedema coma may vary from disorientation
DOI:10.4158/EP13460.OR 2013 AACE.

and lethargy to slow mentation, confusion, cognitive dysfunction, minimal responsiveness, or coma. The decompensated neurologic state may be primary, such as from a cerebrovascular event or due to a drug overdose with sedatives or hypnotics; whereas sepsis, hyponatremia, or other metabolic disturbances are secondary events, which may worsen the cognitive function.

Homeostatic dysfunction resulting from thyroid hormone deficiency is generally insufficient to cause myxedema coma, as the body can compensate through neurovascular mechanisms. A triggering event is usually required to overcome the compensatory mechanisms in a hypothyroid patient. [7] Infection, cerebrovascular or cardiovascular events, cold temperature exposure, medications such as amiodarone, beta blockers, lithium, narcotics, sedatives, diuretics, and metabolic derangements are several examples of such insults. [2, 3] Each patient had at least one identifiable precipitating event and the frequency of these events was in concordance with the findings reported in other studies. [3]

Prolonged untreated hypothyroidism coupled with a triggering event may lead to cardiovascular collapse and shock which may not be responsive to vasopressor therapy alone, until thyroid hormone also is administered [8]. Electrocardiographic abnormalities such as bradycardia, low voltage, nonspecific ST wave inversion, QT prolongation, as well as rhythm abormalities may be seen [9]. Hypotension was commonly seen in our myxedema coma cases, and the frequency of electrocardiographic abnormalities was similar to that reported in the literature [3].

DOI:10.4158/EP13460.OR 2013 AACE.

An impaired ventilatory response and a need for mechanical ventilation are common manifestations in patients with myxedema coma. Decreased respiratory center sensitivity to hypercarbia and hypoxemia may lead to hypoventilation, which may be aggravated further by impaired respiratory muscle function, obesity, and other obstructive processes of the airway such as macroglossia, myxedema of the larynx and nasopharynx, intrinsic processes such as pneumonia, reduced lung volumes, or extrinsic compressive processes such as pleural effusions [1, 10, 11].

Reduced glomerular filtration rate (GFR) in hypothyroid patients is a result of decreased renal plasma flow withwater retention and hyponatremia usually being concomitant findings in these patients [12]. Fluid extravasation, resulting from altered vascular permeability, may present as effusions, nonpitting edema and anasarca. Effects of profound thyroid hormone deficiency on the gastrointestinal system may include decreased intestinal motility with constipation and may progress to paralytic ileus with a quiet and distended abdomen, anorexia, nausea and abdominal pain [23]. In our patients, the metabolic abnormalities occurred with relative equal frequencies but independent of each other, suggesting the importance of appreciation of the multisystemic basis for development of myxedema coma.

The ultimate diagnosis of myxedema coma should be made with biochemical evidence of low levels of serum free T4 and T3, and elevated TSH in patients with primary hypothyroidism, whereas in secondary hypothyroidism the biochemical diagnosis should rely on low, or normal TSH, and low free T4 and total T3 hormone levels and evidence of pituitary dysfunction. None of our patients had biochemical evidence of secondary hypothyroidism.
DOI:10.4158/EP13460.OR 2013 AACE.

Particular attention should be given to patients with biochemical evidence of secondary hypothyroidism that could be difficult to distinguish from the sick euthyroid state. The latter entity represents a physiologic adaptive response of the thyrotropic feedback control to severe illness, and is reflected by biochemical evidence of normal, low, or slightly elevated TSH, depending of the severity of the illness, and low free T4 and T3. Therefore, in order to avoid misclassifying patients with sick euthyroid syndrome as having myxedema coma in the setting of commonly present multiorgan dysfunction, we suggest that appropriate diagnosis of secondary hypothyroidism should be done first, either from history of hypothalamic-pituitary dysfunction, or through imaging studies reflecting organic hypothalamic, or pituitary disease.

This study is limited by virtue of its retrospective design and relatively small sample size, which precluded accurate comparison between groups due to lack of statistical power. Also, due to insufficient published data in all the case reports of myxedema coma assessed from literature, it was not possible to fully validate the scoring system. However, the score demonstrated to have positive predictive value and a high discriminative power.

In conclusion, considering the complex, multisystemic manifestations of hypothyroidism in patients with myxedema coma and the high mortality associated with delays in therapy, a practical guide to earlier diagnosis could be of value. We propose a diagnostic scoring system for myxedema coma based upon data from restrospective cases diagnosed at our institutions, as well as from selected case reports culled from the literature. This scoring system assessed an array of the diagnostic features associated with myxedema coma and found a similar frequency of findings in our cohort of patients as in those assessed from the literature [2, 3, 5].
DOI:10.4158/EP13460.OR 2013 AACE.

This scoring system should be considered in the clinical context of the patient. Further large prospective, well controlled studies are needed to confirm the current findings, and to inform whether such a diagnostic approach to patients with myxedema coma will enable earlier recognition and more effective treatment of this potentially fatal endocrine emergency.

Disclosure Summary: The authors have nothing to disclose.

References:

1. Klubo-Gwiezdzinska J, Wartofsky L. Thyroid emergencies. Endocrinol Metab Clin North Am. 2012; 96:385-403. 2. Rodriguez I, Fluiters E, Prez-Mndez LF, Luna R, Pramo C, Garca-Mayor RV. Factors associated with mortality of patients with myxoedema coma: prospective study in 11 cases treated in a single institution. J Endocrinol. 2004;180:347-350. 3. Dutta P, Bhansali A, Masoodi SR, Bhadada S, Sharma N, Rajput R. Predictors of outcome in myxoedema coma: a study from a tertiary care center. Crit Care. 2008; 12:R1 4. Jordan RM. Myxedema coma. Pathophysiology, therapy, and factors affecting prognosis. Med Clin North Am. 1995;79:185-194 5. Yamamoto T, Fukuyama J, Fujiyoshi A. Factors associated with mortality of myxedema coma: report of eight cases and literature survey. Thyroid. 1999; 9:1167-1174 6. Nicoloff JT. Thyroid storm and myxedema coma. Med Clin North Am. 1985;69:10051017
DOI:10.4158/EP13460.OR 2013 AACE.

7. Fliers E, Wiersinga WM. Myxedema coma. Rev Endocr Metab Disord. 2003;4:137-141 8. Gardner DG. Endocrine emergencies, in D.G. Gardner and D. Shoback, eds. Greenspans Basic and Clinical Endocrinology, McGraw-Hill, New York, NY, USA, 8th edition, 2007. 9. Polikar R, Burger AG, Scherrer U, Nicod P. The thyroid and heart. Circulation. 1993; 87:1435-1441 10. Zwillich CW, Pierson DJ, Hofeldt FD, Lufkin EG, Weil JV. Ventilatory control in myxedema and hypothyroidism. N Engl J Med. 1975;292:662-665 11. Ladenson PW, Goldenheim PD, Ridgway EC. Prediction and reversal of blunted ventilatory responsiveness in patients with hypothyroidism. Am J Med. 1988;84:877-883 12. Montenegro J, Gonzalez O, Saracho R, Aguirre R, Martinez I. Changes in renal function in primary hypothyroidism. Am J Kidney Dis. 1996;27:195-198 13. Kogan A, Kassif Y, Shadel M, Shwarz Y, Lavee J, Or J, Raanani E. Severe hypothermia in myxoedema coma: a rewarming by extracorporeal circulation. Emerg Med Australas. 2011; 23:773-775 14. G Pearse S, D Dahdal M, Grocott-Mason R, W Dubrey S. Myxoedematous pre-coma and heart failure. Br J Hosp Med (Lond). 2011;72:52-53 15. Mallipedhi A, Vali H, Okosieme O. Myxedema coma in a patient with subclinical hypothyroidism. Thyroid. 2011;21:87-89 16. Chu M, Seltzer TF. Myxedema coma induced by ingestion of raw bok choy. N Engl J Med. 2010; 362:1945-1946 17. Chen SY, Kao PC, Lin ZZ, Chiang WC, Fang CC. Sunitinib-induced myxedema coma. Am J Emerg Med. 2009;27:370.e1-370.e3
DOI:10.4158/EP13460.OR 2013 AACE.

18. Cappelli C, Stanga B, Paini A, Gandossi E, Cumetti D, Castellano M, et al. Myxoedema coma precipitated by diabetic ketoacidosis and neuroleptic drugs: case report in an intensive care unit. Intern Emerg Med. 2007;2:147-149 19. Sheu CC, Cheng MH, Tsai JR, Hwang JJ. Myxedema coma: a well-known but unfamiliar medical emergency. Thyroid. 2007;17:371-372 20. Yu CH, Stovel R, Fox S. Chorea--an unusual manifestation in a woman recovering from myxedema coma. Endocr Pract. 2012;18:e43-e48 21. Ahn JY, Kwon HS, Ahn HC, Sohn YD. A case of myxedema coma presenting as a brain stem infarct in a 74-year-old Korean woman. J Korean Med Sci. 2010;25:13941397 22. Kargili A, Turgut FH, Karakurt F, Kasapoglu B, Kanbay M, Akcay A. A forgotten but important risk factor for severe hyponatremia: myxedema coma. Clinics (Sao Paulo). 2010; 65:447-448 23. Yanamandra U, Kotwal N, Menon A, Nair V. Ogilvies syndrome in a case of myxedema coma. Indian J. Endocrinol Metab. 2012;16:447-449

DOI:10.4158/EP13460.OR 2013 AACE.

Table 1. Frequency of events in 21 patients with and without myxedema coma presenting between 1989 2009 at MWHC and VA Medical Center, Washington DC MC n (%) Patients Gender Male Female Age (mean SD ) Date of admission (Nov - Feb) History of hypothyroidism Hypothermia (T < 35o C) Central nervous system Somnolence/lethargy Obtunded Stupor Coma Cardiovascular system Bradycardia (HR < 60) Hypotension Prolonged QT Non-specific ST-T changes Low voltage complexes Bundle branch blocks Pericardial effusion CXR findings Cardiomegaly Pleural effusions 5 (36) 5 (36) 3 (43) 2 (29) 1.0000 1.0000 5 (36) 7 (50) 3 (21) 3 (21) 1 (7) 1 (7) 1 (7) 2 (29) 1 (14) 1 (14) 0 (0) 0 (0) 0 (0) 0 (0) 1.0000 0.1736 1.0000 0.5211 1.0000 1.0000 1.0000 5 (36) 4 (29) 1 (7) 4 (29) 1 (14) 1 (14) 2 (29) 0 (0) 0.6126 0.6244 0.2474 0.2550 8 (57) 6 (43) 68 15 6 (43) 12 (86) 7 (50) 3 (43) 4 (57) 66 23 3 (43) 4 (57) 2 (29) 0.8120 1.0000 0.2800 0.6424 0.6594 14 Non-MC n (%) 7 p-value

Pulmonary edema Pulmonary infiltrates Gastrointestinal symptoms Anorexia, abdominal pain, constipation Decreased bowel sounds Distended, quiet abdomen Metabolic disturbances Decrease in GFR Hypoxemia Hypercarbia Hyponatremia Hypoglycemia Precipitating event Infection Medication non-compliance Furosemide use Cold exposure Medications Hypoglycemia Gastrointestinal bleed Congestive heart failure Hypercapnia Cerebrovascular event Treatment Levothyroxine IV with Steroids Levothyroxine IV without Steroids Levothyroxine PO

3 (21) 2 (14)

3 (43) 2 (29)

0.3544 0.5743

2 (14)

2 (29)

0.5743

2 (14) 1 (7)

0 (0) 0 (0)

0.5333 1.0000

6 (43) 5 (36) 5 (36) 5 (36) 4 (29)

1 (14) 2 (29) 2 (29) 0 (0) 0 (0)

0.3371 1.0000 1.0000 0.1235 0.2550

5 (36) 4 (29) 4 (29) 4 (29) 3* (21) 2 (14) 2 (14) 2 (14) 1 (7) 1(7)

4 (57) 3 (43) 1 (14) 1 (14) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0)

0.3972 0.6514 0.6244 0.6244 0.5211 0.5333 0.5333 0.5333 1.0000 1.0000

9 (64) 3 (21)

1 (14) 0 (0)

0.0635 0.5211

1 (7)

6 (86)

0.0009

Amiodarone (n=2), Amitriptyline (n=1)

SD, standard deviation; T, temperature; HR, heart rate; CXR, chest X Ray; GFR, glomerular filtration rate.

Table 2: Features and variables in 14 patients with myxedema coma (page 1/2)
Patient Age Gender History Cold Tempera NeuroPrecipitating events TSH Free T4 Total T3 (ng/dL)

of hypo- season thyroidism (NovFeb) Yes

-ture (C) cognition

(mU/L) (ng/dL)

49

Yes

33.3

Obtunded

Hypoglycemia Cold exposure

53.4

0.68

50.6

67

Yes

No

36.4

Coma

Infection (PNA) Hypercarbia

28.6

0.59

56.3

84

Yes

No

33.6

Coma

Infection (UTI) GI bleed

125

< 0.3

4 5

41 76

F M

Yes No

No Yes

36.4 36.2

Lethargic Obtunded

Amitriptyline Infection (UTI) Amiodarone Cold exposure

122 170

0.56 0.49 66.3

6 7 8

82 67 49

F F F

Yes Yes Yes

No Yes Yes

36.3 36.3 37

Lethargic Obtunded Lethargic

Infection (UTI) Hypoglycemia GI bleed Furosemide

71 326 57

< 0.2 0.39 0.42

< 40 < 40 < 40

74

Yes

Yes

34.4

Coma

Amiodarone

45

0.2

Cold exposure 10 65 M Yes No 35 Coma CHF Furosemide 11 64 M Yes No 35 Lethargic ? (died at presentation) 12 89 M No Yes 33.8 Stupor CHF Furosemide Cold exposure 13 83 F Yes No 34.4 Obtunded Infection (PNA, UTI) 14 61 M Yes No 36.9 Lethargic/ CVA seizures *TFTs (thyroid function tests) obtained 1 month prior SI conversion factors: To convert freeT4 to nmol/L, multiply by 12.8717; to convert total T3 to pmol/L, multiply by 15.361 PNA, pneumonia; UTI, urinary tract infection; CVA, cerebrovascular accident, CHF, congestive heart failure; GI bleed, gastrointestinal bleed Furosemide 107 0.44 41.2 116 0.59 84 0.3 128.8 0.9 58* 0.6*

Table 2: Features and variables in 14 patients with myxedema coma (page 2/2)
PaHeart Hypo- Hypo- Hyper- Mecha tension xemia carbia -nical ventila -tion 1 87 Yes No No No Sodium (mEq/ L) 137 Glucose Change in GFR EKG findings CXR findings GI symptoms Score

tient rate

(mg/d () L) 42 no QT prolong. No

Decreased 90 intestinal motility

65

No

No

Yes

Yes

104

147

Yes (35)

No

Pleural effusion Infiltrates

Decreased 95 intestinal motility No No 70 95

3 4

62 130

No Yes

No No

No No

Yes No

146 138

50 58

Yes (9) Yes (64)

No No

No Pleural effusions Pulmonary edema

54

No

No

No

No

132

102

Yes (27)

No

No

Constipation

60

6 7

59 83

Yes Yes

Yes Yes

Yes No

Yes Yes

142 133

88 <20

Yes (13) No

No QT pro-

Cardiomegaly No Cardiomegaly No

95 105

long.

Pulmonary edema

61

No

No

No

No

133

81

Yes (19)

No

Cardiomegaly No Pleural effusions

65

70

Yes

Yes

No

No

135

109

N/A

No

cardiomegaly

Abdomin al pain

100

10 11 12

56 46 61

No Yes No

No No No

No No Yes

Yes Yes No

136 133 156

135 71 128

N/A N/A N/A

No No No

No No Pleural effusions

Ileus No No

90 80 60

13

67

Yes

Yes

Yes

Yes

145

175

Yes (15)

QT prolong.

Pleural effusions Pulmonary edema Infiltrates

No

120

14

56

No

No

No

No

138

145

No

No

Cardiomegaly No

75

GFR, glomerular filtration rate; CXR, chest X Ray Heart rate in beats/min; GFR in mL/min.

Table 3: Features and variables in 7 patients without myxedema coma (page 1/2)
Patient Age Gen- History der Cold Tempe NeuroPrecipitating events TSH Free T4 Total T3 (ng/dL)

of hypo- season thyroidism (NovFeb) Yes

-rature cognitio (C) n

(mU/L) (ng/dL)

32

No

31.3

Lethargic

Infecion (bacteremia)

5.67

0.62

56.2

2 3 4 5 6 7

73 52 77 94 45 90

M F F F F M

No Yes Yes Yes Yes No

No Yes No No No Yes

36.8 37 37 36.6 36.6 34.4

Stupor Normal Obtunded Normal Normal Stupor

Infection (PNA) 5.83 Non-compliance 80.6 Non-compliance 9.0 Infection (UTI) 7.2

1.06 0.39 1.3 2.03 0.28 1.4 70.8

Non-compliance 145 Infection (PNA) 11.9 Cold exposure Furosemide

PNA, pneumonia; UTI, urinary tract infection

Table 3: Features and variables in 7 patients without myxedema coma (page 2/2)
PaHeart HypoHypo- Hyper- Mecha -nical ventilation 1 50 No No N/A No 140 75 No(on HD) 2 85 No Yes No Yes 137 80 No (on HD) QT prolong. No Pleural effusions, infiltrates 3 87 No No No No 140 263 No N/A Cardiomegaly 4 5 102 72 No No Yes No Yes No No No 145 144 86 96 No No No No No Cardiomegaly 6 57 No No No No 140 127 No No No consti pation 7 72 Yes No Yes No 145 80 Yes (15) Atrial flutter Cardiomegaly, pleural N/V/c onstipation 100 25 No No 45 25 No 25 No 50 N/A No 50 Sodium Glucose Change EKG findings CXR findings GI symptoms Score

tient rate

tension xemia carbia

(mEq/L) (mg/dL) in GFR ( )

effusions, infiltrates

Heart rate in beats/min; GFR in mL/min; EKG, electrocardiogram; CXR, chest X Ray; GI, gastrointestinal; , delta; HD, hemodialysis.

Table 4. Diagnostic Scoring System for Myxedema Coma Termoregulatory dysfunction (Temperature, oC) >35 32-35 <32 Central Nervous System Effects Absent Somnolent/Lethargy Obtunded Stupor Coma/Seizures Gatrointestinal findings Anorexia/abdominal pain/constipation Decreased intestinal motility Paralytic ileus Precipitating event Absent Present 0 10 5 15 20 0 10 15 20 30 0 10 20 Cardiovascular dysfunction Bradycardia Absent 50-59 40-49 <40 Other EKG changes* Pericardial/pleural effusions Pulmonary edema Cardiomegaly Hypotension Metabolic disturbances Hyponatremia Hypoglycemia Hypoxemia Hypercarbia Decrease in GFR 10 10 10 10 10 0 10 20 30 10 10 15 15 20

*Other EKG changes: QT prolongation, or low voltage complexes, or bundle branch blocks, or non-specific ST-T changes, or heart blocks. A score of 60 or higher is highly suggestive/diagnostic of myxedema coma; a score of 25 -59 is suggestive of risk for myxedema coma, and a score below 25 is unlikely to represent myxedema coma.

Table 5: Features and variables in 22 patients from literature diagnosed with myxedema coma
Ref Pt Age Gen- Temp der (C) Neurocognition Precipitating events Concomitant disorder Heart MAP Rate EKG changes Hypo- Hyperxemia carbia Sodium TSH Free T4 Score

(mEq/L) (mU/L) (ng/dL)

84

34.5

Obtunded

Urinary infection

Pleural effusion

39

110

N/A

No

N/A

133

51.3

0.46

85

75

34.4

Coma

Pneumonia , sepsis

Anemia, DIC, ARDS, septic shock

124

108

N/A

Yes

N/A

122

0.43

0.25

90

70

33.9

Coma

Abdominal surgery

Respiratory 38 failure, shock

115

N/A

Yes

N/A

144

71

0.18

110

65

34.9

Obtunded

Urinary infection

Pericardial effusions Anemia, pneumonia

104

74

N/A

No

N/A

124

2.4

0.23

55

20

34.2

Obtunded

Typhoid fever

114

72

N/A

No

N/A

128

76.04

0.28

45

81

34.8

Coma

Ileus

Respiratory 38 failure, pleural effusion, shock

68

N/A

Yes

N/A

126

28

0.17

130

63

35.0

Obtunded

Urinary infection

Anemia, respiratory failure

124

88

N/A

No

N/A

110

38

0.15

55

83

35.0

Coma

Urinary infection

None

65

95

N/A

No

N/A

122

60.6

0.15

60

79

34.8

Obtunded

Respiratory None infection

52

128

N/A

No

N/A

120

153

0.15

55

10

47

34.9

Obtunded

Urinary infection

Anemia, Respiratory failure

144

112

N/A

No

N/A

126

9.85

0.37

55

11

82

33.6

Obtunded

Pneumonia

Respiratory 38 failure, shock

80

N/A

Yes

N/A

120

78.2

0.5

105

13

12

84

30.0

Global amnesia

N/A

33

60

N/A

No

No

135

63.2

0.17

85

14

13 62

35.3

Delayed response

Non-

Pleural

50

74

Low volt

N/A

N/A

134

>60

undetec 60 table

compliance effusions None Pericardial effusion 88 73

15

14

47

33.2

Lethargic

None

N/A

N/A

Low

6.09

0.83

40>80

16

15

88

36.1

Lethargic

Bok Choy

58

119

N/A

Yes

Yes

132

74.4

undetec 60 table

17

16

68

29.1

Changes in MS

Sunitinib

46

107

N/A

No

No

115

41.4

undetec 75 table

18

17

27

36.6

Changes in MS

Diabetic ketoacidosi s

None

40

98

Low volt

N/A

N/A

132

48

0.4

45

19

18

64

30.1

Changes in MS

Urinary infection

None

60

84

N/A

No

Yes

138

> 200

<0.35

70

20

19 33

35

Coma

Non-

Hypoglyce

50

76

N/A

No

N/A

138

>100

0.24

60

compliance mia 21 20 74 F 34.8 Stupor CVA 59 50 Low Yes No 121 30.12 0.05 100

volt Prol QT 22 21 78 M 35.5 Coma N/A Hypoactive 52 BS 23 22 60 F 37.7 Altered sensorium (obtunded) sepsis Ogilvies syndrome (ileus) bradic 125 Juncti onal rythm N/A N/A 122 341.57 1.6* 75 70 N/A N/A Yes 106 61.24 <0.3 75

Ref, reference; Pt, patient; Temp, temperature; MAP, mean arterial pressure; EKG, electrocardiogram; N/A, not available. DIC, disseminated intravascular coagulation; ARDS, acute respiratory distress syndrome; MS, mental status; CVA, cerebrovascular accident; BS, bowel sounds; Heart rate in beats/min. *Total T4: 1.6 ug/dL (5.6 13.7ug/dL)

1.00

Area under ROC curve = 0.8827

60 65

50

45

25

0.75

70 75 80 90 95

0.25

Sensitivity 0.50

100 105 120

0.00

>120

0.00

0.25

0.50 1 - Specificity

0.75

1.00

a Figure 1. ROC curve of the scoring system for myxedema coma