Case Reports  |   September 1999
Shivering Complicating the Treatment of Neurologically Impaired Surgical and Intensive Care Unit Patients 
Author Affiliations & Notes
  • Mark T. Keegan, M.B., M.R.C.P.I.
  • Frank W. Sharbrough, M.D.
  • William L. Lanier, M.D.
  • *Resident, Department of Anesthesiology. †Emeritus Professor, Department of Neurology (Electroencephalography Laboratory). ‡Professor, Department of Anesthesiology.
Article Information
Case Reports
Case Reports   |   September 1999
Shivering Complicating the Treatment of Neurologically Impaired Surgical and Intensive Care Unit Patients 
Anesthesiology 9 1999, Vol.91, 874. doi:
Anesthesiology 9 1999, Vol.91, 874. doi:
NEW-ONSET rhythmic movement can confound diagnosis and complicate management in patients with central nervous system disease. Rhythmic movement may result from the nonconvulsive effects of drugs (e.g.  , etomidate), the convulsive effects of drugs (e.g.  , local and general anesthetics), epileptic seizures, or other causes. Regardless of the source, inadvertent patient motion may complicate awake neurosurgical procedures, especially if the surgery involves real-time assessment of neurologic function or the acquisition of fine electrical recordings.
Because movement disorders may reflect new or evolving brain injury or result from more benign causes, the identification of the source of the movement may lead to widely divergent patient care interventions. We describe two patients who experienced new-onset, nonconvulsive, rhythmic movement that interfered with their treatment. In both, the source of the movement was normothermic or near-normothermic shivering.
Case Reports 
Case 1
A 29-yr-old man with a history of aqueductal stenosis was admitted to the neurologic intensive care unit after surgery to remove an infected right-sided ventriculoperitoneal shunt. Six hours later, generalized tonoclonic convulsions occurred, and electroencephalography (EEG) confirmed status epilepticus. When initial therapy with diazepam, phenytoin, and phenobarbital failed to control the seizure activity, an amobarbital infusion was begun. The trachea was intubated, and the lungs were mechanically ventilated. Additionally, the patient was treated with 100 mg intravenous phenytoin four times per day.
After 18 h of amobarbital infusion, subsequent EEG was obtained while the patient was without seizures or rhythmical movement. This revealed 0.5–1 Hz δ activity with superimposed rhythmic 10–12 Hz α activity. Also observed were sporadic or quasiperiodic sharp waves arising from the right frontocentral region. These findings were consistent with residual barbiturate effect plus potentially epileptogenic components. The amobarbital infusion rate was reduced to permit further evaluation of the patient's neurologic status. However, over the ensuing 4 h, the patient began experiencing rhythmic movement interpreted as clinical evidence for return of seizures. The amobarbital infusion was returned to the previous delivery rate, and a prolonged EEG recording was begun. The EEG data were consistent with general anesthesia; no seizure activity was observed.
A second attempt to wean the patient 30 h after the original administration of the barbiturate infusion resulted in intense, violent, rhythmic muscle activity, although the patient remained unresponsive to stimuli. The EEG findings were consistent with altered consciousness as a result of residual anesthetic; no clear seizure activity was identified. Instead, EEG tracings contained a rhythmic artifact that was represented in the scalp electrode recordings as 5- to 6-Hz waxing and waning deflections, consistent with a nonepileptic “movement disorder.” The patient's core (bladder) temperature was 37.8°C. Based on the available data, a tentative diagnosis of near-normothermic shivering was made.
The patient was stripped of all clothing. The skin was irradiated using two 500 W infrared heat lamps (Emerson Equipment Model 96H; Emerson Electric, St. Louis, MO), with the light source 1 m from the skin surface. Within approximately 90 s, all rhythmic muscle activity ceased. Two minutes later, the light therapy was discontinued, and the rhythmical activity returned. When this sequence was repeated 5 min later with the same results, the diagnosis of near-normothermic shivering was confirmed. Throughout the test, the patient's temperature was unchanged and no antipyretics were administered.
The amobarbital infusion was not reinitiated. Over the next 24 h, the patient awakened from the amobarbital anesthetic without EEG or clinical evidence of seizures. The trachea was extubated 2 days later. During this period, the patient's skin was kept warm by the use of warm blankets, and the rhythmic muscle activity did not return. The patient was discharged from the intensive care unit 1 week after admission. He left the hospital, without evidence of new neurologic deficit, 10 days later.
Case 2 
A 67-yr-old man was scheduled for awake mini craniotomy and stereotactic placement of a left thalamic stimulator to treat an essential tremor. After headframe placement and magnetic resonance imaging (MRI), the patient was sedated with 1 mg intravenous midazolam, and the scalp was infiltrated with 8 ml of a mixture of 1% lidocaine and 0.375% bupivacaine local anesthetics. The scalp was incised and surgical probes were inserted through burr holes. Using stereotactic techniques, a recording probe was directed into the brain, and electrophysiologic monitoring and clinical examination were used to identify the location of the thalamic tremor generator. During this activity, crescendo and coarse, rhythmic motor activity involving the entire body developed, without alteration in consciousness. As the movement persisted, the patient became extremely anxious and expressed a strong desire to discontinue the surgery. The possibility of discontinuation of surgery also was expressed by the surgeon and neurologist because the new movement prevented the high-quality electrophysiologic recording necessary for successful completion of the surgery. In addition, there was concern for probe- and pinion-induced bleeding.
An extensive consoling conversation with the patient was initiated and a low-dose propofol infusion was begun (approximately 20 μg · kg−1· min−1). This attenuated, but did not abolish, the anxiety, and the patient's underlying tremor and new-onset movements persisted. Based on these observations, a diagnosis of normothermic shivering was considered.
A convective warming blanket was placed next to the patient's skin, from the chest downward, and this was connected to a forced-air warmer (Bair Hugger Polar Air Model 600; Augustine Medical, Inc., Eden Prairie, MN) set at 43°C. All rhythmic motion (in excess of the patient's baseline hand tremor) ceased within 3 min. Concomitantly, the patient's anxiety level decreased still further. Surgery proceeded uneventfully, and the site for stimulator electrode placement was established. After 5 h of surgery, general anesthesia was induced to allow placement of the stimulator pulse generator. Temperature, as measured by esophageal stethoscope, was 36.0°C at this time, and the temperature remained unchanged at the end of the procedure 90 min later. Activation of the stimulator greatly attenuated the abnormal tremor activity on the right side. The patient was discharged on the first postoperative day.
Five months later, a fractured stimulator wire necessitated repeat surgery. The patient was sedated with an infusion of intravenous propofol at 50 μg · kg−1· min−1, a stereotactic headframe was placed, and computed tomography was performed. No efforts were initiated to keep the patient warm during transportation and scanning. At return to the operating room, a urinary bladder catheter with temperature-sensing capability (Bardex 400-series; CR Bard, Covington, GA) was inserted. The patient's temperature was 35.6°C, with an ambient operating room temperature of 18°C. He complained of feeling cold. On this occasion, in the setting of the background tremor only (i.e.  , without visible evidence of shivering), the patient also experienced mild anxiety. However, the anxiety was far less than that experienced during the initial surgery. A forced-air warmer (temperature set at 38°C) was used again, and, despite a core temperature that remained at 35.6°C, the patient's anxiety ceased within minutes. Surgery proceeded uneventfully and postoperative recovery was rapid.
We describe two patients whose clinical course was complicated by shivering. In the first, the shivering interfered with the clinical and electrophysiologic diagnosis of seizures, thus extending the period in which the patient received an amobarbital anesthetic. In the second, intraoperative shivering again precluded accurate electrophysiologic monitoring and, additionally, increased the risk of bleeding. This patient experienced extreme anxiety associated with the shivering. Had the coupled shivering and anxiety not been treated, it would have been necessary to discontinue the surgery.
In patient 1, we had evidence that the shivering occurred during a period without core hypothermia. (A possible cause is that the hypothalamic “set point” for temperature regulation was increased, perhaps because a fever was developing. 1) In both patients, shivering ceased within minutes after initiating skin warming. Thus, the period of warming, that was sufficient to stop shivering, was insufficient to meaningfully alter core temperature.
Distinguishing between shivering and other forms of rhythmic motor activity can be difficult. One possible approach is to characterize and quantify the frequency of the motor activity and compare it to the reported properties of motor activity in different physiologic states. Electromyographic recordings at frequencies of 5–12 Hz can represent shivering, but there is overlap with a variety of physiologic and pathologic conditions. 2,3 Although shivering also may exhibit a fairly characteristic 4- to 8-cycles/min waxing and waning pattern, 4 this was not specifically identified as an artifact within the EEG in the first patient we described.
The effect of barbiturate infusion on the first patient's rhythmic movements was believed to represent evidence for underlying seizure activity. In hindsight, it probably represented the influence of an anesthetic agent on the “interthreshold range”5 for thermoregulation (i.e.  , the 0.2°C range outside of which thermoregulatory mechanisms are initiated). For the same reason, the early use of propofol probably prevented shivering with hypothermia (i.e.  , bladder temperature 35.6°C) during the second patient's second surgery. 6 After we suspected a diagnosis of normothermic or near-normothermic shivering in our patients, the treatment was patterned after the reports of Sharkey et al.  7,8 This therapy is based on the principle that reduction in skin temperature, independent of core hypothermia, can initiate shivering. Warming the skin in such patients will abruptly halt shivering. Sharkey et al.  7 irradiated shivering patients with a warming light. The patients ceased shivering in 61 ± 10 s (mean ± standard deviation). However, after the warming light irradiation was discontinued, the patients resumed shivering in 43 ± 7 s. We used this principle of skin warming to effectively treat shivering in our patients and, in patient 1, to establish a diagnosis. In patient 2, direct skin irradiation was impossible because of space limitations caused by surgical draping and equipment. Instead, we used a forced-air warming blanket. 9,10 In our patients, the rapid onset of the warming effect (patients 1 and 2) and the rapid offset (patient 1) were valuable in establishing the origin of the undesired movement.
Patient 2 had bouts of anxiety during both surgeries. In the first, when the patient shivered violently, the anxiety was debilitating. Although not relieved by the conventional approaches of conversation and sedation, the anxiety was halted by skin warming. Although core hypothermia plus input from thermal sensors within the skin may have directly affected the anxiety spells, the experience during the first surgery also suggests a correlation between muscle activity and the anxiety. Surface warming, which rapidly halted the shivering, also abolished the anxiety. These observations suggest that shivering per se,  perhaps acting through muscle receptors (e.g.  , muscle spindles), served to modulate the anxiety attack. Such an interpretation is consistent with a large body of experimental evidence that shows that increased muscle afferent traffic (as would be expected during shivering) has the potential to desynchronize the EEG and produce alterations in mentation and behavior in awake and lightly anesthetized subjects. 11 
In summary, we present two cases in which the presence of shivering confounded the diagnosis and treatment of patients having central nervous system disease. Warming the skin, without altering core temperature, halted the shivering within minutes and, in one patient, also alleviated anxiety. Based on this experience, when patients are normothermic, we recommend the inclusion of shivering in the differential diagnosis of patients who experience new-onset rhythmic movements. Additionally, our report confirms the effectiveness of using heating lamps or forced-air warmers to prevent, diagnose, and treat shivering.
Guyton AC, Hall JE: Textbook of Medical Physiology, 9th Edition. Philadelphia, WB Saunders, 1996, pp 911–22
Pozos RS, Iaizzo PA: Shivering and pathological and physiological clonic oscillations of the human ankle. J Appl Physiol 1991; 71: 1929–32
Horowitz BA: Homeostatic responses to acute cold exposure: Thermogenic responses in birds and mammals, Handbook of Physiology: Environmental Physiology. Edited by Fregly MJ, Blatteis CM. New York, Oxford University Press, 1996, pp 359–77
Sessler DI: Temperature monitoring, Anesthesia, 4th Edition. Edited by Miller RD. New York, Churchill Livingstone, 1994, pp 1363–82
Lopez M, Sessler DI, Walter K, Emerick T, Ozaki M: Rate and gender dependence of the sweating, vasoconstriction, and shivering thresholds in humans. A NESTHESIOLOGY 1994; 80: 780–8
Matsukawa T, Kurz A, Sessler DI, Bjorksten AR, Merrifield B, Cheng C: Propofol linearly reduces the core temperature thresholds for vasoconstriction and shivering thresholds. A NESTHESIOLOGY 1995; 82: 1169–80
Sharkey A, Lipton JM, Murphy MT, Giesecke AH: Inhibition of postanesthetic shivering with radiant heat. A NESTHESIOLOGY 1987; 66: 249–52
Sharkey A, Gulden RH, Lipton JM, Giesecke AH: Effect of radiant heat on the metabolic cost of postoperative shivering. Br J Anaesth 1993; 70: 449–50
Lennon RL, Hosking MP, Conover MA, Perkins WJ: Evaluation of a forced-air system for warming hypothermic postoperative patients. Anesth Analg 1990; 70: 424–7
Cheng C, Matsukawa T, Sessler DI, Ozaki M, Kurz A, Merrifield B, Lin H, Olofsson P: Increasing mean skin temperature linearly reduces the core-temperature thresholds for vasoconstriction and shivering in humans. A NESTHESIOLOGY 1995; 82: 1160–8
Lanier WL, Iaizzo PA, Milde JH, Sharbrough FW: Cerebral and systemic effects of movement in response to a noxious stimulus in lightly anesthetized dogs. Possible modulation of cerebral function by muscle afferents. A NESTHESIOLOGY 1994; 80: 392–401