CLINICIANS have observed that patients with carotid disease or mass lesions of the brain may awaken from general anesthesia with a focal neurologic deficit that resolves rapidly in the postoperative period. The temporal profile of this recovery suggests a pharmacologic etiology, and the term "differential awakening" has been coined recently to describe this phenomenon. 
The phenomenon itself never has been characterized carefully, however, and evidence that anesthetics are involved has been entirely anecdotal. [2-4]
Therefore, the purpose of this study was to test the hypothesis that sedatives can worsen or unmask focal neurologic deficits in patients with underlying neurologic dysfunction.
Materials and Methods
Fifty-four patients scheduled for elective surgery to remove a supratentorial mass lesion of the brain or scheduled for carotid endarterectomy for cerebrovascular disease were studied in the immediate preoperative period, with institutional approval and after informed consent was obtained. Subjects were aged 18 yr or older and were of ASA physical status 1-3. Among patients scheduled for removal of mass lesions of the brain, particular effort was made to enroll those with lesions in the fronto-parietal regions, because these patients were assumed to be at greater risk of having or developing focal motor deficits. All patients with carotid disease were considered suitable candidates for enrollment. Patients were excluded only if they were unable to comprehend and cooperate with a neurologic examination because of impaired mental status during enrollment.
Patients were studied in the induction room preoperatively. They received no preanesthetic medication. Pulse oximetric arterial oxygen saturation was monitored continuously and supplemental oxygen administered as indicated. Blood pressure was monitored intermittently with an automated device (Propaq 106, Protocol System, Beaverton, OR). A baseline neurologic examination, including assessment of mental status, strength, and reflexes (Table 1
) was performed by an investigator who knew the nature of the surgery but not the details of the patients' symptoms or baseline neurologic examination. Furthermore, to control for the potential confounding influence of sedative-induced inattention or lack of cooperation, a series of questions and commands designed to assess orientation, attention, language, comprehension and memory was included. 
Once the baseline neurologic examination was complete, patients received either midazolam (Hoffman-Roche, Nutley, NJ) or fentanyl (Abbot Laboratories, North Chicago, IL) intravenously in divided doses to achieve a predetermined endpoint of a lightly sedated but fully cooperative patient, and the neurologic examination was repeated 5 min later. If there was no change in sedation or the neurologic examination after the first dose, additional doses were administered and a third neurologic examination performed 5 min later.
A positive finding was defined as a unilateral, focal deterioration in strength or performance and/or the development of abnormal reflexes relative to the patient's baseline neurologic examination that was not associated with a concomitant deterioration in mental status, blood pressure, or oxygen saturation. Data (i.e., motor deterioration vs. no deterioration) were analyzed using Fisher's Exact Test.
Twenty-seven patients with a supratentorial mass lesion of the brain and 27 with carotid disease were enrolled in the study. Fifty-two percent were men and 48% were women, and the average age was 56 yr (range 23-85 yr). All subjects completed the study as designed, and no patient was eliminated after enrollment because of untoward effects or excessive sedation. Thirty-one patients received midazolam and 23 received fentanyl. The total doses (mean+/-SD) of midazolam and fentanyl were 2.8 +/-1.3 mg (range, 1-8 mg) and 170+/-60 micro gram (range, 75-300 micro gram), respectively. Ninety-four percent of patients received a total of 1-4 mg midazolam or 100-200 micro gram fentanyl. Three patients were given larger dosages because they remained fully awake and alert and had no change in the neurologic examination. No patient developed an arterial hemoglobin oxygen saturation of less than 95% or a decrease in blood pressure more than 15% from baseline or a systolic blood pressure of less than 100 mmHg.
The overall incidence of a deterioration in the neurologic examination after sedation was 30%. Thirty-five percent of the patients who received midazolam had significant changes in the neurologic examination, whereas 22% had positive findings after administration of fentanyl. Patients who exhibited changes in the neurologic examination were nearly evenly divided between those with mass lesions (9 of 27) and those with carotid disease (7 of 27).
Sedation-induced changes in the neurologic examination ranged from subtle to profound. Subtle changes occurred in 11 patients; 6 developed a new unilateral pronator drift alone, whereas 5 had mild worsening of weakness accompanied by either an increase in clumsiness (finger tapping) or development of unilateral hyperreflexia, ankle clonus, or Babinski response. Profound deterioration in the neurologic examination occurred in five patients; in each, focal paresis progressed to complete plegia. Although we did not formally study the time course of this phenomenon, it appeared that sedative-induced deterioration in motor function had partially or completely resolved before the induction of general anesthesia 10-20 min later.
To test the hypothesis that a preexisting deficit predisposes to sedative-induced worsening of neurologic function, data were reanalyzed according to whether the patient had a history of a motor deficit at some time in the course of the illness. Of the 54 patients studied, 13 had a motor deficit at the time of the study and 9 had a remote history of a focal motor deficit from which they had recovered. Of these 22 patients, 14 received midazolam, whereas 8 were given fentanyl. Among the 32 patients with no preexisting motor weakness or a history of weakness, none experienced new symptoms after administration of midazolam (17 patients) or fentanyl (15 patients). In contrast, among those with a history of focal motor deficits, 11 of 14 (79%; P < 0.001) and 5 of 8 (63%; P < 0.02) who received midazolam or fentanyl, respectively, demonstrated exacerbation or unmasking of neurologic dysfunction relative to the baseline examination (Figure 1
). Despite the fact that the pathophysiology of the underlying diseases was different, there were no differences in response to sedation between patients with a history of a deficit caused by tumor or carotid disease (9 of 10 vs. 7 of 12, P = 0.16 with Fisher's Exact Test). In all cases, the sedative-induced motor deficit occurred in the same distribution as the prior or existing motor deficit.
Collectively, therefore, 73% of patients (16 of 22) with a history of a motor deficit had a significant deterioration after administration of sedatives. All 13 patients who had onset of a focal deficit within 3 months before surgery had deterioration of neurologic function with sedation, whereas only 33% of those with a longstanding deficit (i.e., onset greater than 3 months earlier) had such changes (Figure 2
). On subsequent review of the medical records of patients with carotid disease and a history of a motor deficit, 10 of 12 had preoperative computed tomography or magnetic resonance imaging scans available. These 10 had radiographic evidence of cerebral infarct(s), and all 12 were diagnosed clinically as having had strokes.
Although it was postulated that anesthetic agents can produce focal neurologic dysfunction in some patients, and anecdotal evidence to support this view is plentiful, [1-4]
there has been no systematic study of this phenomenon in humans. Using the preoperative period to avoid the confounding influence of surgery, this study demonstrates that midazolam and fentanyl can exacerbate or unmask focal motor deficits in patients with prior neurologic dysfunction. These drug-induced changes in the neurologic examination frequently were subtle and, although we did not formally assess the duration of the effect, appeared to be transient. These drug effects occurred predominantly in patients with a history of a focal motor deficit that was either present at the time of the study or had resolved in the recent past. Therefore, this phenomenon seems to be prevalent in patients with prior acute or subacute focal neurologic symptoms but evidently does not occur de novo in patients without a history of prior dysfunction. This held true even in a patient with a functional-magnetic resonance imaging documented mass lesion directly abutting the motor cortex; he, however, had no preexisting motor deficit and no sedative-induced changes in his motor examination.
This study was conducted in the preoperative period to eliminate the potential influences of surgery and multiple sedative/anesthetic medications. Nevertheless, the study had limitations. First, although we used an extensive mental status examination to ensure that observed changes in the neurologic examination were not due to inattention or lack of cooperation, we cannot entirely exclude the possibility that patient motivation was reduced by sedation. To the extent that signs such as pronator drift, development of ankle clonus, hyperreflexia, and a Babinski response are independent of attentiveness or cooperation, this seems unlikely. Second, because the examiners were not blinded to the drug administered or to the type of surgery, there was potential for examiner bias. Any potential bias was mitigated, however, because during the study the examiners were unaware of the specific nature of the patient's disease (e.g., left- or right-sided pathology) or associated neurologic dysfunction. Third, there was no attempt to pharmacologically reverse a drug-induced change in the neurologic examination because we did not believe it ethically or medically appropriate to antagonize sedation precipitously in patients about to have major surgery. Fourth, we preferentially included patients with mass lesions in the fronto-parietal regions of the brain and, as such, may have selected patients who would demonstrate neurologic deterioration after administration of midazolam or fentanyl. Finally, the study was limited to formal examination of reflexes and motor function because these parts of the neurologic examination are the most objective and easiest to quantify. Because we did not quantify or include changes in other components of the neurologic examination, it is conceivable that the incidence of sedative-induced dysfunction reported here is an underestimate. Indeed, a profound but transient recurrence of an expressive aphasia developed in one patient after administration of midazolam. The patient was not included in our positive group because she failed to develop motor or reflex changes.
There are several potential mechanisms for the effects reported here. First, there may be altered drug uptake, distribution, and/or metabolism in diseased or abnormal areas of the brain. Luxury perfusion may develop after a cerebral infarction and may persist for as long as 6 weeks. 
Such pathophysiology may account for the greater incidence of neurologic changes after sedation in patients with focal motor deficits of less than 3 months duration. Second, neurons in and around abnormal brain may be more sensitive to sedatives. Support for this hypothesis comes from evidence that the pharmacodynamics of pentobarbital is altered in brain-injured rats such that lower concentrations of pentobarbital are required to prevent movement in response to tail clamping. 
The third proposal is based on the premise that some of the redundancy in neuronal numbers or circuitry that characterizes the normal state is lost focally in diseased brain. Despite such limited reserve, even minimal impairment of the remaining functioning neurons by a centrally acting drug may produce an exaggerated response. Lastly, functional and structural reorganization may occur as an adaptation to neural injury.
Support for this concept is illustrated in one functional positron-emission tomography scan study that demonstrated the development of new motor pathways in patients recovered from ischemic strokes. 
This theory of brain "plasticity" could explain why the patients with longstanding prior motor deficits who had more time to develop different motor pathway connections were less likely to experience a sedative-induced motor deficit. It is important to recognize that no data speak directly to these issues, and it is not clear whether any of these proposed mechanisms apply. However, the ability of two agents with very different receptor affinities and pharmacology to exacerbate or unmask neurologic dysfunction suggests that the phenomenon is not specific to a particular class of sedative/hypnotics but may be a more general property of centrally acting compounds.
This study confirms that drug-related worsening of a neurologic deficit occurs, particularly among patients with acute or subacute focal motor neurologic dysfunction. Although this study examined patients only in the perioperative period, it is reasonable to assume that transient focal neurologic deficits in the early postoperative period also may be drug related. Consequently, the possibility of a pharmacologic effect should be added to the differential diagnosis of a seemingly new focal motor deficit in the immediate postoperative period.
1. Cucchiara RF: Differential awakening. Anesth Analg 1992; 75:467.
2. Miller RA, Crosby G, Sundaram P: Exacerbated spinal neurologic deficit during sedation a patient with cervical spondylosis. ANESTHESIOLOGY 1987; 67:844-6.
3. Benzel EC, Hadden TA, Nossaman BD, Lancon J, Kesterson L: Does sufentanil exacerbate marginal neurologic dysfunction? J Neurosurg Anesthesiol 1990; 2:50-2.
4. Baskin DS, Hosobuchi Y: Naloxone reversal of ischemic neurologic deficits in man. Lancet 1981; 2:272-5.
5. Folstein SE, McHugh PR: "Mini-mental state": A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975; 12:189-98.
6. Ramsey RG: Neuroradiology with computed tomography, Arteriosclerosis. Edited by Ramsey RG. Philadelphia, WB Saunders, 1981, p 316.
7. Archer DP, Priddy RE, Tang TKK, Sabourin MA, Samanani N: The influence of cryogenic brain injury on the pharmacodynamic of pentobarbital. ANESTHESIOLOGY 1991; 75:634-9.
8. Chollet F, Di Piero V, Wise RJS, Books DJ, Dolan RJ, Frackowiak RSJ: The functional anatomy of motor recovery after stroke in humans: A study with positron emission tomography. Ann Neurol 1991; 29:63-71.
© 1996 American Society of Anesthesiologists, Inc.