Secondary Logo

Journal Logo

Anesthetic Pharmacology: Research Report

A Randomized, Prospective, Double-Blinded Study of Physostigmine to Prevent Sedation-Induced Ventilatory Arrhythmias

Karan, Suzanne B. MD*; Rackovsky, Elia BA*; Voter, William A. MA*; Kanel, Jason A. MD*; Farris, Nick BS*; Jensen, Joshua BS*; Liu, Lynn MD; Ward, Denham S. MD, PhD*‡

Author Information
doi: 10.1213/ANE.0000000000000834
  • Free


Morbidity and mortality associated with sedation, even when administered by qualified personnel, is a critical clinical problem.1,2 Soto et al.3 found that 26% of patients undergoing sedation, with a variety of drugs, had apneic episodes of 20 seconds that were not detected by the anesthetists. Sedatives, typically benzodiazepines and opioids, are frequently used in combination for patient comfort during procedures and have the possible side effects of causing respiratory depression through apneas, hypoventilation, and/or upper airway collapse.4–6 Sedation-related events are reported as high as 41% in some cohorts and pose a serious risk for patients predisposed to respiratory events, such as those with obstructive sleep apnea (OSA).7,8 This is a serious concern which can ultimately be lethal and, in routine cases, may interrupt and prolong the procedure.

Hedner et al.9 demonstrated that physostigmine reduced the incidence of sleep-disordered breathing, although it may not be as effective as first reported.10 In addition, prior studies have demonstrated excitatory effects of physostigmine on the respiratory system,11–14 supporting the hypothesis that physostigmine may reduce apneas (obstructive and/or central) and hypopneas during sedation. The goal of this study was to determine whether pretreatment with physostigmine decreases ventilatory arrhythmias during moderate sedation with midazolam and remifentanil in healthy male volunteers. The term “ventilatory arrhythmia” has been used to describe disordered breathing during experimental conditions by other investigators.15–17 Ventilatory arrhythmias will be quantified by the measurement of the sedation apnea–hypopnea index (S-AHI) as the primary outcome and nadir oxygen saturation as the secondary outcome. Both of these outcomes are clinically relevant in moderate sedation. Measurement of the hypercapnic ventilatory response (HCVR) and arousal state (electroencephalogram [EEG] and bispectral index [BIS]) is performed to elucidate possible mechanisms. For example, we hypothesized that if physostigmine reversed sedation-related ventilatory arrhythmias, one possible explanation might be supported by an awake-appearing EEG in the presence of a clinically sedated subject. This pilot study was designed to see whether physostigmine could be useful as a concurrent treatment during moderate (conscious) sedation.


Figure 1
Figure 1:
Timeline for participant preparation and measurements.

The study was registered in Clinical (Identifier NCT01171118) before enrollment of subjects. Written informed consent, approved by our IRB, was obtained from all participants. We prospectively recruited 10 males, aged 18 to 45 years, body mass index (BMI) <25, to participate in this 2-day, randomized, prospective placebo-controlled experiment. Recruitment was accomplished with the assistance of flyers posted around our university. Subjects were free of medical or psychiatric illness. Before participation, subjects were assessed with a baseline electrocardiogram, examined to verify normal physical anatomy, and had their airways evaluated (Mallampati classification, presence of retrognathia/prognathia, thyromental distance). Weight, height, and BMI were quantified, and all subjects were asked to complete the Epworth Sleepiness Scale (ESS).18 The ESS rates, on a 4-point scale (0–3), the chances of dozing off or falling asleep in 8 different scenarios. The severity of OSA has been correlated with an ESS over 12; thus, we used this survey to help identify subjects who may have an undiagnosed sleep disturbance. Once consented, 2 study days were scheduled to begin at 7:00 AM, 1 day each for an infusion of physostigmine and of placebo (normal saline). Figure 1 shows the study protocol and timeline. S1 and S2 refer to measurement points before and after the sedation period, respectively.


All participants were monitored with pulse oximetry (model 3900, Datex-Ohmeda Inc., Madison, WI) and noninvasive blood pressure (Datascope 2100, Datascope Corp., Paramus, NJ). Blood pressure was taken every 5 minutes during sedation and every 15 minutes during recovery. Oxygen via nasal cannula was initially administered at 2 L/min while sedation was being titrated to effect. Respiratory inductance plethysmography (RIP) (Respitrace, Ambulatory Monitoring Inc., Ardsley, NY) was used to measure the excursions of the abdominal and chest walls, and oronasal airflow thermistors (Embla® BreathSensor, Broomfield, CO) were used to identify and characterize reduced or absent airflow. Processed (BIS, Model A-2000, Aspect Medical, Natick, MA) and unprocessed EEG (Embla® N7000, Broomfield, CO) were measured in the frontal, cortical, temporal, and occipital regions. Two IV lines were inserted for the administration of physostigmine/placebo separately from the sedation infusion (remifentanil and midazolam).

Study Drug

Physostigmine has rapid pharmacokinetics and is best given as an IV infusion to ensure a constant plasma concentration. In sleep apnea subjects, Hedner et al.9 used 0.12 µg·kg−1·min−1 (approximately 0.5 mg/h in a 70-kg patient) to achieve a plasma concentration of 3 to 4 nmol/L. Furey et al.19 infused 1 mg/h and found steady-state concentrations after 40 minutes; Beilin et al.20 used the same infusion rate (1 mg/h) preceded by a bolus of 1.5 mg. Previous studies had explored IV bolus administration of physostigmine (0.5–2 mg),21 but due to a report of very rapid plasma elimination, the administration of an infusion, rather than a single bolus, was thought to be most effective for obtaining a steady state for the experimental procedure.22 We used the same dose as Beilin et al. of 0.24 µg·kg−1·min−1 (1 mg/h in a 70-kg person) without a bolus infusion to ensure an adequate plasma level without the side effects that might have been produced by bolus dosing. An hour of infusion time was established before experimental measurements were performed to obtain steady-state concentrations. We did not measure plasma concentrations, because these have been reported previously in similar subjects. The infusions of physostigmine and placebo were administered in double-blind fashion on the 2 study days with the physostigmine diluted in normal saline by the research pharmacist.

Sedation Protocol

Midazolam was chosen, because it is the most commonly used IV anxiolytic for moderate sedation.23 Although remifentanil is rarely used by sedation practitioners, we chose it for its rapid pharmacokinetics which facilitated earlier discharge of our volunteers. A combination of the subjective Observer’s Assessment of Alertness/Sedation (OAA/S) score,24 BIS monitor, and a thermal pain score was used to measure a constant level of sedation/analgesia. First, the midazolam infusion was set for an effect-site target of 100 ng/mL25 (Standard Syringe Pump 22, Harvard Apparatus, Holliston, MA) and was adjusted individually for a BIS of 70 to 80 and an OAA/S score of 2 to 3. The remifentanil infusion was then started at 0.05 μg·kg–1·min–1 and titrated upward to provide a decrease in pain by 3 points (see below). The infusion rate of remifentanil was consistent with a target effect-site concentration of 0.2 to 0.3 ng/mL.26

The analgesic effect was tested using thermode application (Precision Pain Source/PPS-3, Cygnus, Patterson, NJ) after presensitization of the subject’s forearm with capsaicin cream.27 In brief, before the initiation of the drug infusions (including the placebo/physostigmine), the temperature of the thermode was adjusted to achieve a pain score of 5 to 6 of 10. The thermode site was presensitized with 0.075% capsaicin cream.28 The remifentanil infusion was adjusted to achieve a pain score of 2 to 3 points less than baseline.

Hypercapnic Ventilatory Response

For determining the HCVR, subjects were fitted with a facemask and air flow was measured with a bidirectional turbine flowmeter (VMM-2B Ventilation Measurement Module and Flow Cartridge Assembly, Adult, Interface Associates, Ventura, CA) and airway gases were sampled continuously by a combined infrared (CO2) and fast paramagnetic (O2) analyzer (Gemini Respiratory Monitor; CWE, Inc., Ardmore, PA). The dynamic end-tidal forcing technique29 was used to measure the HCVRs at 3 time points during the study (Fig. 1). In brief, this entailed rapidly changing the PETCO2 to obtain 3 or 4 response points (average of 1 minute of hypercapnic breathing following a 2- to 3-minute equilibrium at the desired increased PETCO2, usually 50 and 55 mm Hg, but the exact levels were determined by the starting PETCO2). All responses were done in normoxia. The HCVR was quantified by linear regression to report the ventilation at a PETCO2 of 60 mm Hg (HCVR60) as a composite measure of the HCVR.

Measurement of S-AHI

We defined the S-AHI as the sum of the number of respiratory events, normalized by the total time period (S-AHI = [central apneas + obstructive apneas + hypopneas + mixed events]/h). After the drugs were titrated to the desired sedation/analgesia level, the participant was monitored while in the supine position for 2 hours, with the subject’s head placed on a Shea headrest to ensure consistency of head position between experimental runs.30 The S-AHI was assessed for 1 hour at each of 2 inspired oxygen concentrations (order randomly assigned by coin toss)—room air (0.21 FiO2) or with nasal cannula O2 at 2 to 3 L/min (approximately 0.35 FiO2). Although traditional sleep laboratory AHI quantifications are performed in room air (and thus the nadir SpO2 is measured), we chose to add the arm occurring during the administration of oxygen as would mimic typical clinical sedation practice.

The respiratory events were classified strictly on the basis of respiratory effort measured by the RIP and flow as measured by the nasal thermistor. Sleep state, EEG arousal, and oxygen saturation were assessed, but not taken into account, when identifying respiratory events. Although traditional sleep studies adjust the measured AHI by sleep stage, we quantified the AHI during the hour-long time intervals and then describe the presence of various sleep stages during that time without adjusting the AHI measurement. The use of standard sleep-scoring methods was applied to quantify the sedation-related apnea–hypopnea index with the recognition that significant sleep-related respiratory events may not be clinically relevant in the sedation realm.

Central apnea was defined as events during which there was no airflow and neither thoracic nor abdominal movement for >10 seconds. Obstructive apnea was defined as events during which there was no airflow, but there was continued inspiratory effort (normal or exaggerated) with the abdominal and thoracic movements out of phase for >10 seconds. Hypopnea was defined as events during which there is an episodic reduction in air flow and the magnitude of the RIP tracings (>30% for >10 seconds) without a phase shift. Hypopneas were also classified as central or obstructive (paradoxical breathing was present). An event was called “mixed” if there was any intermingling of the above events in any episode.31

Measurement of EEG-Quantified “Sleep” Stages

Figure 2
Figure 2:
Example of EEG scoring. The characteristics of N2 “sleep”—spindles, K-complexes, and vertex waves—are present during this sedation EEG. Sleep spindles, which are thought to be generated in the thalamic reticular nucleus, are bursts of oscillatory brain activity characterized by 12–14 Hz and lasting for at least a half second. K-complexes are brief high amplitude delta frequency waves that help to characterize slow wave sleep. Both of these waveforms are thought to assist in memory consolidation. Vertex waves are sharper than the typical waveform and also characteristic of stage 2 sleep.
Figure 3
Figure 3:
Sedated subject, with physostigmine, on room air. Subject begins in N1 “sleep” (the first stage of sleep) undergoes an arousal and then resumes N1 “sleep.” Before arousal, subject has an obstructive apneic event which is manifested by the lack of flow on oronasal thermistor and “paradoxical” chest and abdominal movement. When the airway is obstructed, the chest and the abdomen are out of phase from each other to increase negative pressure to overcome the obstruction. After self-arousal, subject resumes normal airflow and synchronous breathing (thermistor and thoracic/abdominal belt waveforms align). The SpO2 declines with obstruction and increases to baseline after arousal.

Despite the participants maintaining a constant level of sedation and analgesia (as measured by the OAA/S, BIS, pain scores, and clinical observation), EEG analysis allowed for a more granular resolution of the sedation periods into EEG-quantified “sleep” stages. (“Sleep” is encapsulated in single quotes to designate to the reader that sedation sleep may not represent physiologic sleep.) EEG staging from awake and N1 and N2 “sleep” was accomplished by using standardized criteria (Figs. 2 and 3).32 In staging sleep, N1 is typically defined as the first stage of sleep and is characterized by a low voltage and mixed frequency pattern. The second stage of sleep, N2, is usually signaled by sleep spindles and/or K complexes. The periods of sedation sleep that were measured in our study did not encompass deeper stages such as N3 and/or rapid eye movement sleep. The staging of the EEG “sleep” state in each of the epochs was accomplished blindly by one of the authors. The “sleep” efficiency was calculated as the percentage of time the subject was in N1 and N2 “sleep” during the 1 hour of recorded time. Normal sleep efficiency ranges between 80% and 90% of total sleep time. The BIS value was recorded every 5 minutes throughout the sedation periods.

Statistical Methods and Data Analysis

Sample Size Considerations

There are no directly relevant studies on which to base the sample size for this pilot study. In a study conducted by Hsu et al.,33 8 male subjects were sedated with either remifentanil or dexmedetomidine and were assessed for ventilatory arrhythmias. When sedated with remifentanil, 4 of 7 subjects exhibited S-AHIs above 15. Although a higher dose of remifentanil was used in that study, since we were using the combination benzodiazepine and opioid, we hypothesized a synergistic effect on ventilatory arrhythmias as was seen by Bailey et al.4 using midazolam and fentanyl. These data suggest that clinically significant S-AHI effects can be observed (within subjects) relative to sedation and that 10 subjects would be a reasonable sample size.

The primary outcome variable is the S-AHI, and pairwise comparisons were made between placebo versus physostigmine under either room air or O2. The pairwise differences were tested for deviation from normality with Shapiro-Wilk W test. Nonnormal differences were tested with the Wilcoxon paired signed rank test (exact method). Statistical significance was set at P < 0.05 for both tests. Secondary outcomes are nadir SpO2, BIS, sleep efficiency, and percent time spent “awake” and in N1 and N2 “sleep.” For the primary and secondary outcomes, the significant P values are reported for both the parametric and nonparametric tests for significance (PT and PSR are the P values for the signed rank and t test, respectively). The experimental conditions (total drug doses, pain scores, and the HCVR60) were tested for differences paired t tests, and the probability of deviation from normality is given as PSW. The HCVR60 at the time points S1 and S2 time points (see Fig. 1) were compared with the control HCVR60 using a paired t test with the significance level set at P < 0.025 (Bonferroni correction). Data are given as mean ± SD unless otherwise noted. Confidence intervals (95% CIs) were calculated using both the bias corrected, nonparametric bootstrap method (1000 iterations) and under normality assumptions (Student t test). CIBS and CIT are the confidence intervals calculated using the bootstrap method or normally, respectively. The STATA (Stata Corp, College Station, TX) software package was used for all statistical analysis.


Ten males completed the study (Table 1), there were no adverse events related to the study, and all subjects tolerated the experimental procedures. One subject experienced nausea and vomiting after the end of a study day during which physostigmine was administered. The administration of physostigmine did not necessitate significant changes in the administration of remifentanil or midazolam to maintain constant analgesia or sedation (Table 1). The targets for OAA/S, decrease in pain score, and BIS were achieved in all subjects (Table 1). Although the BIS value was lower with physostigmine in both room air and oxygen, this did not reach statistical significance.

Table 1
Table 1:
Study Outcome Data, Participant Demographics, and Drug Doses

Sedation during placebo infusion resulted in a clinically significant S-AHI (>15)32 in only 3 of the 10 subjects (Fig. 4). Two subjects experienced a clinically significant S-AHI on both room air and O2 and one on only O2. The mean S-AHI tended to be reduced in both conditions (13.4 ± 18.8 [mean ± SEM] [CIT = −29.0 to 55.9; CIBS = −18.0 to 50.45], PT = 0.4, PSR = 0.96 in oxygen vs. 6.2 ± 8.0 [mean ± SEM] [CIT = −12.2 to 24.7; CIBS = −5.2 to 23.8], PT = 0.46, PSR = 0.96 in room air) (Table 1). While neither difference was statistically significant, there was clinically significant variation between subjects (Fig. 4). Physostigmine reduced the S-AHI in all 5 instances of a clinically significant S-AHI on placebo (from 72.8 ± 44.2 to 5.8 ± 11.3 events/h, PT = 0.0393, PSR = 0.0431, delta = 67.0 ± 22.2 [mean ± SEM] [CIT = 5.3–128.6; CIBS = 29.2–111.7]). Interestingly, during placebo/O2 2 participants had very high S-AHIs (Fig. 4). As expected for participants with a low BMI and no history of OSA, most of the ventilatory arrhythmias under study conditions were of central origin (Table 2), but obstructive events did occur (Fig. 3). Most of the pure obstructive events occurred with placebo/O2. The nadirs in oxygen saturation were not different in either room air or with supplemental oxygen.

Table 2
Table 2:
Total Counts of the Number of Events (Obstructive, Central, Mixed Obstructive, and Central)
Figure 4
Figure 4:
Effects of physostigmine on the sedation apnea–hypopnea index (S-AHI) during (A) room air and (B) oxygen. The colors are for the same subjects in each panel. The dark green line is from the subject with Epworth Sleepiness Scale score of 15. The left panel contains data from 2 subjects who had S-AHIs of 0 for both conditions—their lines are thus not distinct.

Although by ordinary clinical measures (OAA/S, pain score, and BIS) the subjects were at a constant level of sedation and analgesia whether receiving placebo or physostigmine infusions, a detailed analysis of the EEG revealed some more subtle differences. One set of EEG data could not be analyzed for technical reasons for placebo–room air, and one set of BIS data for physostigmine (room air and oxygen) was unavailable for analysis. Although there was a general tendency for there to be less EEG scored “sleep,” this did not always reach statistical significance. Nonetheless, with supplemental O2, the “sleep” efficiency (%) was decreased by 14.2 ± 3.3 (mean ± SEM) (PT = 0.0019, PSR = 0.0069, CIBS = 8.2–20.5, CIT = 6.8–21.6), and the percent time “awake” increased by 10.5 ± 3.8 (mean ± SEM) (PT = 0.0229, PSR = 0.0367, CIBS = 3.8–17.1, CIT = 1.8–19.1). The stage of sleep most affected seemed to be N2 “sleep” (Table 1).

Changes in the chemical control of breathing can have an effect on ventilatory arrhythmias. The baseline HCVR60 was 3 to 4 times larger than normal average resting ventilation and, as expected, the sedation significantly reduced the HCVR60 at both time points by for placebo and at S1 for physostigmine (Table 1). A few of the HCVR runs could not be analyzed; technical malfunction caused lack of data during 1 run with placebo at S1 and S2, and erratic breathing hampered analysis during 5 other HCVR runs (2 placebo at S1, 1 at S2, and 2 physostigmine at S1). There were also no significant differences between the HCVR60 on the placebo and physostigmine infusion days.


This study was designed to test whether an infusion of physostigmine would prevent or reduce ventilatory arrhythmias, as measured by the S-AHI, during moderate sedation. Although we did not find a significant effect of physostigmine on the S-AHI in this cohort of normal volunteers, there are still several interesting aspects to the study.

Previous studies suggest that altering the cholinergic system may be excitatory in the rat hypoglossal motonuculeus34 and influence airway patency in humans.10 Acetylcholine is a neurotransmitter involved in both sleep and the control of breathing.35 Specific to upper airway patency, acetylcholine has both excitatory and inhibitory effects at the hypoglossal motor nucleus. Stimulation of muscarinic cholinergic receptors at the level of the hypoglossal motor nucleus suppresses pharyngeal motor neurons.36 In contrast, central nicotinic cholinergic receptor activation produces excitability (and may restore wakefulness drive to breathe).37 While some studies have shown the net effect favoring suppression of motor activity,36 this was in rats, and it is possible that in humans the net effect would favor nicotinic-mediated excitation. Although not significant, the increased BIS values and decreased sleep time in the physostigmine arm of this study support a possible mechanism of an increased arousal or “wakefulness” drive. Although the EEG-based measurements may have supported increased “wakefulness” in the physostigmine arm, clinically the subjects all appeared to be at the same level of analgesia and sedation (OAA/S and pain scores).

It is important to note that the ventilatory arrhythmias noted during sedation in normal healthy subjects are potentially different from what would be observed in patients with OSA either during normal sleep or during sedation. Our normal subjects had a predominance of central apneas, as is typical of sedation, particularly with opioids and benzodiazepines. Because of the involvement of the cholinergic system in the neural control of airway patency, physostigmine may have a greater effect when there is a predominance of obstructive apneic episodes. Although the total number of obstructive events in this cohort was low, they were reduced by physostigmine. It is possible that in a group of patients with a propensity for obstructive events during sedation (e.g., patients with OSA), physostigmine could have a significant effect. Interestingly, 1 subject had a clinically significant ESS score of 15, indicating a higher probability of his having OSA. This subject had S-AHIs of 137 and 27.3 on placebo, reduced to 1.2 and 0 for O2 and room air, respectively, during the physostigmine. This particular subject inspires further study into the application of this drug for the treatment of sedation-related ventilatory arrhythmias in at-risk individuals, rather than for its preventative use in unselected patients.

Clinically, physostigmine is most often used in the postoperative period to reverse emergent delirium caused by drugs with a central anticholinergic effect.38 Its properties as an antidote to sedation-related (i.e., opioids and/or benzodiazepines) respiratory depression have also been investigated and have been attributed to its ability to promote nonspecific central arousal via cholinergic mechanisms.11 Other studies have found varying effects of physostigmine on the respiratory effects of opioids and benzodiazepines. Snir-Mor et al.39 showed that the depression of the HCVR caused by morphine was reversed by physostigmine, but in this study all subjects were pretreated with droperidol. Bourke et al.40 found that physostigmine did not antagonize the depressed HCVR caused by morphine and, in the same study, did not find any significant respiratory depression caused by diazepam alone. After a bolus injection of diazepam, Spaulding et al.41 found that physostigmine caused further depression of the HCVR. We found a significant depression of the HCVR by the combination of remifentanil and midazolam that was not immediately reversed by physostigmine, although there was a tendency for partial reversal by the end of the protocol (see Table 1, HCVR results). The partial reversal of ventilatory depression might also represent some acute tolerance to the effects of the remifentanil.

This measurement of the S-AHI with time-synched EEG recordings and apnea severity scores provides a temporal link between cortical arousability and ventilatory stability. BIS recordings were used as commonly used by anesthesiologists to quantify sedation objectively. However, during dynamic changes in cortical activity (such as during titration of rapid- and short-acting sedative agents), real-time EEG recording provides a more robust measurement than the BIS score.42 This is important in interpreting data from different studies that measure effects of sedation on upper airway tone and ventilatory stability.43 Sleep scientists have recently been delving into the relationship between respiratory arousals and cortical activity in an effort to understand phenotypic variability among patients with OSA.44 Since sedation blunts chemical and cortical-wakefulness drives, the measurement of the S-AHI mandates the use of an EEG classification of sedation states (as is done in sleep studies) so that additional studies can be compared with one another.

Limitations and Future Studies

There are several mechanisms by which physostigmine might improve the S-AHI. We accounted for the possible effect of physostigmine on sedation and analgesia by adjusting the infusions of midazolam and remifentanil to maintain a constant level of sedation and analgesia. Higher doses of both midazolam and remifentanil were required, but the increase did not reach statistical significance. Our rationale for this study was to see whether physostigmine could be used as a preventative treatment for respiratory events during sedation as would be typically performed in the gastrointestinal endoscopy clinic. In this situation, it was envisioned that physostigmine would be started at the same time as the sedation and would not require waiting for an event that would interrupt the procedure. In this application, the sedation provider would titrate the sedation drugs to achieve the desired effect. That is, if physostigmine only reduced the level of sedation, thereby preventing the ventilatory arrhythmias, it would not be useful. This protocol aimed to incorporate clinical relevance and applicability. Because we were interested in using physostigmine to decrease ventilatory arrhythmias during sedation, our study design resulted in the use of placebo and physostigmine on different days. There could be significant variation in the severity of the ventilatory arrhythmias on different days in the same subjects, even with the same level of sedation.

An obvious limitation to this study is the lack of statistical power to detect a positive outcome. This occurred mainly because our sedation protocol did not incur as many respiratory events as have been observed in studies of similar healthy cohorts. Bailey et al.4 used boluses of midazolam and fentanyl and observed most of the episodes occurring within 5 minutes of the drug administration, most likely reflecting the higher peak drug levels. We chose to study conditions during a more steady-state level of sedation by infusing the drugs instead of administering boluses, since boluses of opioids have themselves been shown to cause transitional breathing.45 Conceivably, there could also be differences between the central effects of fentanyl and remifentanil. Mortazavi et al.46 found that, in cats anesthetized with halothane, fentanyl decreased acetylcholine release (as measured by microdialysis) in the medial pontine reticular formation but remifentanil did not. Since this area of the pons is known to have sleep- and activation-modulating effects, it is conceivable that our results are unique to remifentanil.

If the subject-to-subject variability remains the same, it would require over 100 subjects to make the differences we found statistically significant. Since such a large study would not be practical, the results found in this small cohort are important in designing a study that would better define the usefulness of physostigmine in treating ventilatory arrhythmias. Since 5 of the 6 instances of a clinically significant S-AHI (Fig. 4) improved with physostigmine, a protocol that used physostigmine to treat a clinically significant S-AHI might have shown significant results. We also used healthy volunteer participants, all with a normal BMI and ESS (except for 1 subject as noted above). Patients with an increased BMI and/or OSA are at greater risk for ventilatory arrhythmias during sedation, and the effects of physostigmine may be different in this cohort.

The use of standard sleep scoring metrics in the sedation realm may overestimate the incidence of “events” that are actually clinically relevant. On the other hand, expanding the time required to classify an apnea from 10 to 30 seconds might underestimate the drug effects. With the application of oxygen obviating nadir saturation as an outcome in 1 arm of the study, we sought a heightened granularity in ventilatory measurements. There is still a dearth of studies quantifying, comparing, and contrasting sleep- and sedation-related polysomnography. Thus, and in accordance with other investigators,47 we sought to further bolster the “apples to apples” comparisons by standardizing the AHI quantification.

Finally, the use of EEG to quantify the sleep state during sedation needs further validation. EEG signals that are commonly used to denote sleep stages (i.e., spindles, K-complexes) have also been used to describe drug-induced EEG changes. The similar convergence of sleep and anesthesia/sedation on the EEG is still unclear.48,49 Our use of EEG analysis during sedation raises questions regarding the relationship between physiologic sleep and sedation.


The pharmacologic prevention of sedation-induced ventilatory arrhythmias would improve clinical care both by a possible decrease in mortality and morbidity but also by improvements in sedation efficiency. Our protocol did not show a significant overall treatment effect of IV physostigmine in a healthy cohort. However, as indicated by the effect of physostigmine in reducing clinically significant ventilatory arrhythmias in those few subjects who experienced this problem in our study, future studies using physostigmine to treat ventilatory arrhythmias in susceptible patients may be worthwhile.


Name: Suzanne B. Karan, MD.

Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.

Attestation: Suzanne B. Karan has seen the original study data, reviewed the analysis of the data, approved the final manuscript, and is the author responsible for archiving the study files.

Name: Elia Rackovsky, BA.

Contribution: This author helped design the study, conduct the study, and analyze the data.

Attestation: Elia Rackovsky has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

Name: William A. Voter, MA.

Contribution: This author helped design the study, conduct the study, and analyze the data.

Attestation: William A. Voter has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

Name: Jason A. Kanel, MD.

Contribution: This author helped analyze the data and write the manuscript.

Attestation: Jason A. Kanel has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

Name: Nick Farris, BS.

Contribution: This author helped conduct the study and analyze the data.

Attestation: Nick Farris has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

Name: Joshua Jensen, BS.

Contribution: This author helped conduct the study, analyze the data, and write the manuscript.

Attestation: Joshua Jensen has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

Name: Lynn Liu, MD.

Contribution: This author helped conduct the study, analyze the data, and write the manuscript.

Attestation: Lynn Liu has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

Name: Denham S. Ward, MD, PhD.

Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.

Attestation: Denham S. Ward has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

This manuscript was handled by: Ken B. Johnson, MD.


The authors would like to thank Michael Perlis, PhD, Associate Professor of Psychology and Director of the Behavioral Sleep Medicine Program, University of Pennsylvania, and Wilfred Pigeon, PhD, Associate Professor of Psychiatry, University of Rochester, for their advice in experimental design relating to “sleep” measurements.


1. Bhananker SM, Posner KL, Cheney FW, Caplan RA, Lee LA, Domino KB. Injury and liability associated with monitored anesthesia care: a closed claims analysis. Anesthesiology. 2006;104:228–34
2. Vibert JF, Foutz AS, Caille D, Hugelin A. Respiratory rhythm multistability during sleep-wake states. Brain Res. 1988;448:403–5
3. Soto RG, Fu ES, Vila H Jr, Miguel RV. Capnography accurately detects apnea during monitored anesthesia care. Anesth Analg. 2004;99:379–82
4. Bailey PL, Pace NL, Ashburn MA, Moll JW, East KA, Stanley TH. Frequent hypoxemia and apnea after sedation with midazolam and fentanyl. Anesthesiology. 1990;73:826–30
5. Santiago TV, Edelman NH. Opioids and breathing. J Appl Physiol. 1985;59:1675–85
6. Hines R, Barash PG, Watrous G, O’Connor T. Complications occurring in the postanesthesia care unit: a survey. Anesth Analg. 1992;74:503–9
7. Kaw R, Michota F, Jaffer A, Ghamande S, Auckley D, Golish J. Unrecognized sleep apnea in the surgical patient: implications for the perioperative setting. Chest. 2006;129:198–205
8. Memtsoudis SG, Besculides MC, Mazumdar M. A rude awakening–the perioperative sleep apnea epidemic. N Engl J Med. 2013;368:2352–3
9. Hedner J, Kraiczi H, Peker Y, Murphy P. Reduction of sleep-disordered breathing after physostigmine. Am J Respir Crit Care Med. 2003;168:1246–51
10. Hedner J, Grote L, Zou D. Pharmacological treatment of sleep apnea: current situation and future strategies. Sleep Med Rev. 2008;12:33–47
11. Ebert U, Oertel R, Kirch W. Physostigmine reversal of midazolam-induced electroencephalographic changes in healthy subjects. Clin Pharmacol Ther. 2000;67:538–48
12. Meuret P, Backman SB, Bonhomme V, Plourde G, Fiset P. Physostigmine reverses propofol-induced unconsciousness and attenuation of the auditory steady state response and bispectral index in human volunteers. Anesthesiology. 2000;93:708–17
13. Caldwell CB, Gross JB. Physostigmine reversal of midazolam-induced sedation. Anesthesiology. 1982;57:125–7
14. Berkenbosch A, Olievier CN, Wolsink JG, DeGoede J, Rupreht J. Effects of morphine and physostigmine on the ventilatory response to carbon dioxide. Anesthesiology. 1994;80:1303–10
15. Jacono FJ, Peng YJ, Nethery D, Faress JA, Lee Z, Kern JA, Prabhakar NR. Acute lung injury augments hypoxic ventilatory response in the absence of systemic hypoxemia. J Appl Physiol (1985). 2006;101:1795–802
16. Gonsenhauser I, Wilson CG, Han F, Strohl KP, Dick TE. Strain differences in murine ventilatory behavior persist after urethane anesthesia. J Appl Physiol (1985). 2004;97:888–94
17. Catley DM, Thornton C, Jordan C, Lehane JR, Royston D, Jones JG. Pronounced, episodic oxygen desaturation in the postoperative period: its association with ventilatory pattern and analgesic regimen. Anesthesiology. 1985;63:20–8
18. Johns MW. A new method for measuring daytime sleepiness: the Epworth sleepiness scale. Sleep. 1991;14:540–5
19. Furey ML, Pietrini P, Alexander GE, Mentis MJ, Szczepanik J, Shetty U, Greig NH, Holloway HW, Schapiro MB, Freo U. Time course of pharmacodynamic and pharmacokinetic effects of physostigmine assessed by functional brain imaging in humans. Pharmacol Biochem Behav. 2000;66:475–81
20. Beilin B, Bessler H, Papismedov L, Weinstock M, Shavit Y. Continuous physostigmine combined with morphine-based patient-controlled analgesia in the postoperative period. Acta Anaesthesiol Scand. 2005;49:78–84
21. Wiklund L. Reversal of sedation and respiratory depression after anaesthesia by the combined use of physostigmine and naloxone in neurosurgical patients. Acta Anaesthesiol Scand. 1986;30:374–7
22. Hartvig P, Pettersson E, Wiklund L, Lindström B. Pharmacokinetics and effects of 9-amino-1,2,3,4-tetrahydroacridine in the immediate postoperative period in neurosurgical patients. J Clin Anesth. 1991;3:137–42
23. Kain ZN, Mayes LC, Bell C, Weisman S, Hofstadter MB, Rimar S. Premedication in the United States: a status report. Anesth Analg. 1997;84:427–32
24. Chernik DA, Gillings D, Laine H. Validity and reliability of the observer’s assessment of alertness/sedation scale: study with intravenous midazolam. J Clin Psychopharmacol. 1990;10:244–51
25. Albrecht S, Ihmsen H, Hering W, Geisslinger G, Dingemanse J, Schwilden H, Schüttler J. The effect of age on the pharmacokinetics and pharmacodynamics of midazolam. Clin Pharmacol Ther. 1999;65:630–9
26. Lang E, Kapila A, Shlugman D, Hoke JF, Sebel PS, Glass PS. Reduction of isoflurane minimal alveolar concentration by remifentanil. Anesthesiology. 1996;85:721–8
27. Petersen KL, Jones B, Segredo V, Dahl JB, Rowbotham MC. Effect of remifentanil on pain and secondary hyperalgesia associated with the heat–capsaicin sensitization model in healthy volunteers. Anesthesiology. 2001;94:15–20
28. Karan S, Voter W, Palmer L, Ward DS. Effects of pain and audiovisual stimulation on the opioid-induced depression of the hypoxic ventilatory response. Anesthesiology. 2005;103:384–90
29. Swanson GD, Bellville JW. Step changes in end-tidal CO2: methods and implications. J Appl Physiol. 1975;39:377–85
30. Armstrong JJ, Leigh MS, Sampson DD, Walsh JH, Hillman DR, Eastwood PR. Quantitative upper airway imaging with anatomic optical coherence tomography. Am J Respir Crit Care Med. 2006;173:226–33
31. Carskadon M RAKryger M, Roth T, Dement WC. Monitoring and staging human sleep. In: Principles and Practice of Sleep Medicine. 2006 Philadelphia, PA Elsevier-Saunders:665–82
32. Silber MH, Ancoli-Israel S, Bonnet MH, Chokroverty S, Grigg-Damberger MM, Hirshkowitz M, Kapen S, Keenan SA, Kryger MH, Penzel T, Pressman MR, Iber C. The visual scoring of sleep in adults. J Clin Sleep Med. 2007;3:121–31
33. Hsu YW, Cortinez LI, Robertson KM, Keifer JC, Sum-Ping ST, Moretti EW, Young CC, Wright DR, Macleod DB, Somma J. Dexmedetomidine pharmacodynamics: part I: crossover comparison of the respiratory effects of dexmedetomidine and remifentanil in healthy volunteers. Anesthesiology. 2004;101:1066–76
34. Chamberlin NL, Bocchiaro CM, Greene RW, Feldman JL. Nicotinic excitation of rat hypoglossal motoneurons. Neuroscience. 2002;115:861–70
35. Douglas CL, Demarco GJ, Baghdoyan HA, Lydic R. Pontine and basal forebrain cholinergic interaction: implications for sleep and breathing. Respir Physiol Neurobiol. 2004;143:251–62
36. Liu X, Sood S, Liu H, Horner RL. Opposing muscarinic and nicotinic modulation of hypoglossal motor output to genioglossus muscle in rats in vivo. J Physiol. 2005;565:965–80
37. Schwartz JR, Roth T. Neurophysiology of sleep and wakefulness: basic science and clinical implications. Curr Neuropharmacol. 2008;6:367–78
38. Duvoisin RC, Katz R. Reversal of central anticholinergic syndrome in man by physostigmine. JAMA. 1968;206:1963–5
39. Snir-Mor I, Weinstock M, Davidson JT, Bahar M. Physostigmine antagonizes morphine-induced respiratory depression in human subjects. Anesthesiology. 1983;59:6–9
40. Bourke DL, Rosenberg M, Allen PD. Physostigmine: effectiveness as an antagonist of respiratory depression and psychomotor effects caused by morphine or diazepam. Anesthesiology. 1984;61:523–8
41. Spaulding BC, Choi SD, Gross JB, Apfelbaum JL, Broderson H. The effect of physostigmine on diazepam-induced ventilatory depression: a double-blind study. Anesthesiology. 1984;61:551–4
42. Ozgoren M, Bayazit O, Kocaaslan S, Gokmen N, Oniz A. Brain function assessment in different conscious states. Nonlinear Biomed Phys. 2010;4(suppl 1):S6
43. Litman RS. Upper airway collapsibility: an emerging paradigm for measuring the safety of anesthetic and sedative agents. Anesthesiology. 2005;103:453–4
44. Younes MK. Role of arousals in the pathogenesis of obstructive sleep apnea. Am J Respir Crit Care Med. 2003;169:623–33
45. Ferguson LM, Drummond GB. Acute effects of fentanyl on breathing pattern in anaesthetized subjects. Br J Anaesth. 2006;96:384–90
46. Mortazavi S, Thompson J, Baghdoyan HA, Lydic R. Fentanyl and morphine, but not remifentanil, inhibit acetylcholine release in pontine regions modulating arousal. Anesthesiology. 1999;90:1070–7
47. Salinas-Miranda AA, Shaffer-Hudkins EJ, Bradley-Klug KL, Monroe AD. Student and resident perspectives on professionalism: beliefs, challenges, and suggested teaching strategies. Int J Med Educ. 2014;5:87–94
48. Murphy PJ, Erskine R, Langton JA. The effect of intravenously administered diazepam, midazolam and flumazenil on the sensitivity of upper airway reflexes. Anaesthesia. 1994;49:105–10
49. John ER, Prichep LS. The anesthetic cascade: a theory of how anesthesia suppresses consciousness. Anesthesiology. 2005;102:447–71
© 2015 International Anesthesia Research Society