Elevation of Thyrotropin Upon Accidental Hypothermia in an Elderly Man

Introduction

A lteration of the pituitary-thyroid axis upon accidental hypothermia in human patients is not well recognized. We report here a rare case of elevation of thyrotropin (TSH) associated with accidental hypothermia.

Patient

A 75-year-old man was admitted through the emergency department on a cold January day because of consciousness disturbance.

He had suffered damage to the cervical spinal cord 2 years earlier and had been experiencing difficulties moving his extremities. He had been receiving 20 mg mianserin hydrochloride and 5 mg warfarin as treatment for depression and deep vein thrombosis of the lower extremities, respectively. The mean outside temperature was approximately −2.0°C (28.4°F), but his house was poorly insulated and inadequately heated. Consequently, during the 10 days prior to admission, his body temperature had fallen as low as 34°C (93.2°F). During the 5 days before admission, his speech had become slow and sluggish, and on the day of admission, he was barely able to communicate.

Upon arrival, his level of consciousness was E4 and V4 on the Glasgow Coma Scale: M grading was difficult due to partial tetraplegia. His speech was slow and faint. The patient's height was 183 cm and body weight 75 kg. His axillary temperature was too low to measure and his rectal temperature was 29.5°C (85.1°F). His blood pressure was 113/60 mmHg, pulse rate 40 beats/min, arterial pH 7.332, and SpO2 97% (with room air). Goiter was not palpable, skeletal muscle was atrophic in the extremities, which were rigid, and pitting edema, not myxedema, was present on the trunk and the lower extremities. The general laboratory data showed hypoproteinemia, liver dysfunction, low creatinine, hyperkalemia, and neutrophilia, which were all mild in degree. The international normalized ratio for prothrombin time was significantly elevated at 7.71. Serum TSH was elevated, and free triiodothyronine (FT₃) and free thyroxine (FT₄) lowered, but thyroid-related autoantibodies were all negative (Table 1). Adrenocorticotropic hormone (ACTH) and cortisol were normal. Fasting plasma glucose (FPG) was in the range of nondiabetic hyperglycemia, with elevated immunnoreactive insulin (IRI) and C-peptide immunoreactivity (CPR). Glycated hemoglobin was normal, indicating that the hyperglycemia was a transient phenomenon. A drip infusion of 0.9% saline was initiated and 20 mg menatetrenone was administered iv for hypothermic coagulopathy. After 2 hours of mild extracorporeal warming, rectal temperature rose to 31°C (87.8°F), and a drip infusion of 500 mL 0.5% glucose solution containing 100 mg hydrocortisone was continued for 6 hours: the ACTH, cortisol, IRI, and CPR data were not yet available at this point. Only modest lowering of FT₃ and FT₄, normal creatinine kinase, and total cholesterol, a minimal increase in the cardiothoracic ratio seen in the chest X-ray (57%), and an absence of low voltage in the electrocardiogram and myxedema, taken together suggested that hypothyroidism was not the main cause of hypothermia. Therefore, a small amount of T₄, 50 μg, was administered orally.

By the next morning, body temperature had risen to >36°C (>96.8°F). The patient developed pneumonia and was unable to take oral medication, so thyroxine was withheld. He remained mildly febrile for a week until resolution of the pneumonia and during this time there was no further recurrence of hypothermia. Serum TSH decreased exponentially and the patient's condition had become almost normal by day 22 without the need for thyroid hormone replacement (Table 1). FT₃ and FT₄ were found to be slightly lowered and elevated, respectively, during the same period, in the subnormal range. FPG, IRI, and CPR had become normal by day 2 (Table 1). Serum luteinizing hormone was slightly suppressed for the patient's age on admission, but remained at an appropriate level from day 2 onwards (Table 1). Prolactin was normal and remained at a similar level throughout the observation period (Table 1).

Discussion

Severe hypothyroidism can cause hypothermia and a form of consciousness disturbance known as myxedema coma. However, hypothermia in this patient was thought to be not due to hypothyroidism but due to a combination of the following circumstances. First, the patient's far from ideal housing condition had resulted in prolonged exposure to low environmental temperatures. Second, heat generation by shivering was insufficient due to skeletal muscle atrophy and rigidity caused by a remote neck injury. Third, nutrition was suboptimal, and he had been taking an antipsychotic drug. Body temperature was quickly normalized within a day and kept normal thereafter. In this patient, we found an intriguing change in the functionality of the pituitary-thyroid axis. Serum TSH upon admission was clearly elevated with slightly depressed FT₃ and FT₄. Thereafter, TSH decreased exponentially toward a normal level over a period of 3 weeks. While serum FT₄ increased gradually during the same period, there was a decrease in serum FT₃ and in the FT₃:FT₄ ratio. At the end of the observation period, the patient settled into the state known as “nonthyroidal illness syndrome.” The elevation of TSH was thus considered a response to accidental hypothermia, driven by input from the hypothalamus or higher central nervous system. Although the patient took 50 μg T₄ once, the dosing was too small to explain subsequent changes in the pituitary-thyroid axis. As far as we are aware, cold stress-induced activation of the hypothalamic-pituitary-thyroid axis has been reported in experimental animals (1) but not in patients with accidental hypothermia (2). The failure to detect TSH elevation in one early study might have been due to the low sensitivity of the TSH radioactive iodine immunoassay (2). In a more recent study, extreme lowering of body temperature down to 18°C in anesthetized children receiving cardiac surgery was associated with transient elevation of TSH (3). What distinguishes such cases from that of our patient, however, is that the elevation of TSH took place simultaneously with a significant lowering of FT₃ and FT₄, implying that the rise in TSH was due to physiological feedback regulation from reduced serum thyroid hormone levels (3). The gradual lowering of FT₃ and the FT₃:FT₄ ratio after recovery from hypothermia in our patient indicated that hypothermia, through increased sympathetic tone, had stimulated conversion of T₄ to T₃ (4). TSH elevation may well be a purposeful adaptation to cold, protecting the body, especially the central nervous system, through heat generation. In fact, the biochemical and clinical findings in this patient appear to closely resemble features of “polar T₃ syndrome”. This condition was well defined by Reed and colleagues (5,6) on the basis of studies of dwellers/workers in Antarctica, in whom a subtle reduction of serum T₃ and FT₃ was observed, along with normal or low serum T₄ and FT₄, elevated unstimulated TSH, and exaggerated serum TSH response to TRH. In our case, one of the things to highlight may be the chronic exposure to cold.

In the evaluation of patients with accidental hypothermia, the possibility of an adaptive elevation of TSH should be borne in mind, although this has yet to be described in the literature (3,7). Ours is a single observation in a frail elderly patient. Therefore, this clearly warrants further studies of the adaptation of the pituitary-thyroid axis in patients with accidental hypothermia in the population at large.