ISSN NUMBER: 1938-7172
Issue 5.5

Michael A. Fiedler, PhD, CRNA

Contributing Editors:
Penelope S Benedik, PhD, CRNA, RRT
Mary A Golinski, PhD, CRNA
Gerard Hogan Jr., DNSc, CRNA
Alfred E Lupien, PhD, CRNA
Lisa Osborne, PhD, CRNA
Dennis Spence, PhD, CRNA
Cassy Taylor, DNP, DMP, CRNA
Steven R Wooden, DNP, CRNA

Assistant Editor
Jessica Floyd, BS

A Publication of Lifelong Learning, LLC © Copyright 2011

New health information becomes available constantly. While we strive to provide accurate information, factual and typographical errors may occur. The authors, editors, publisher, and Lifelong Learning, LLC is/are not responsible for any errors or omissions in the information presented. We endeavor to provide accurate information helpful in your clinical practice. Remember, though, that there is a lot of information out there and we are only presenting some of it here. Also, the comments of contributors represent their personal views, colored by their knowledge, understanding, experience, and judgment which may differ from yours. Their comments are written without knowing details of the clinical situation in which you may apply the information. In the end, your clinical decisions should be based upon your best judgment for each specific patient situation. We do not accept responsibility for clinical decisions or outcomes.

Table of Contents

Implications of postoperative visual loss: steep trendelenburg position and effects on intraocular pressure

Anesthesia-related maternal mortality in the united states: 1979-2002

  Polysomnographic variables predictive of adverse respiratory events after pediatric adenotonsillectomy

Low-dose ketamine with multimodal postcesarean delivery analgesia: a randomized controlled trial

Rolapitant for the prevention of postoperative nausea and vomiting: a prospective, double-blinded, placebo-controlled randomized trial

Dreaming in sedation during spinal anesthesia: A comparison of propofol and midazolam infusion

Uncomplicated removal of epidural catheters in 4365 patients with international normalized ratio greater than 1.4 during initiation of warfarin therapy


Attention subscribers licensed in Alabama, Alaska, Idaho, Kentucky, Nevada, and New Mexico:

This issue contains 1 PHARMACOLOGY specific CE credit.

Implications of postoperative visual loss: steep trendelenburg position and effects on intraocular pressure

AANA J 2011;79:115-121

Molley BL


Purpose Describe the relationship between prolonged steep trendelenburg position and intraocular pressure in patients undergoing lower abdominal laparoscopic procedures.


Background At the author’s institution, they had a patient who experienced postoperative visual loss due to posterior ischemic optic neuropathy after a prolonged (7.5 hr) laparoscopic prostatectomy in steep trendelenburg position. The event occurred despite normotension, normal acid base status, and minimal blood loss. It was postulated that surgical positioning combined with the prolonged laparoscopic procedure in steep trendelenburg position, defined as >30 degrees, lead to facial and orbital edema. This edema may then have contributed to a compartment syndrome within the orbit of the eye resulting in significantly elevated intraocular pressure (IOP). The high venous pressure and interstitial edema may have been such that autoregulation of intraocular blood flow was unable to compensate. This could result in decreased ocular perfusion pressure (OPP) and ophthalmic ischemia. In this study OPP = MAP – IOP.


Methodology This study was initially developed as a quality assurance project after the Department of Public Health required monitoring of IOP after the above sentinel event occurred. The project was later turned into a prospective observational study. Data from the quality assurance monitoring was combined with the prospective data. Patients scheduled for prolonged lower abdominal laparoscopic procedures requiring steep trendelenburg position were enrolled. Surgeries included laparoscopic prostatectomy, bowel, and hysterectomy procedures scheduled for at least 120 minutes (unknown which procedures were robotic assisted). Patients were excluded if they had a history of glaucoma, diabetes, history of eye surgery, vascular disease, or malignant hyperthermia.


IOP was monitored with a handheld tonometer (Tono-Pen XL). Anesthesia providers credentialed to perform IOP monitoring recorded the measurements. All patients had an arterial line placed and blood pressure was maintained with vasopressors as needed to keep MAP at approximately 80 mm Hg. The first measurement took place after induction of anesthesia with the patient in the supine position. While the patient was in steep trendelenburg position, IOP was monitored every 30 minutes for the duration of the procedure. A protractor was used to document the degree of steep trendelenburg. Fluids were generally limited unless surgical conditions dictated otherwise.


In the recovery room and on postoperative day one patients were questioned about the presence of eye pain and visual acuity. Descriptive and inferential statistics were used to analyze the results. Repeated measures analysis of variance was used to evaluate changes in IOP over time.


Result Over three years, 43 patients ASA I-III were enrolled. Six patients were excluded because the procedure was returned to the supine position before the impact of steep trendelenburg could be evaluated over time. This left 37 patients in the analysis (n = 21 women and n = 16 men), with a mean age of 50 years (range 31 to 78 years) and mean BMI = 28 kg/m2. In 40.5% (n = 15) of patients the BMI was > 30 kg/m2. The frequency of ASA status was ASA I (11%), ASA II (75%), and ASA III (14%). No data was presented on the frequency of different surgical procedures. Surgeries lasted an average of 180 minutes. Average crystalloid administered was 2,500 mL of lactated Ringer’s solution. Average estimated blood loss was 250 mL (range 50-600 mL).


There was a significant increase in IOP over time (P = 0.001). IOP was significantly higher at 30, 60, 90, and 120 minutes, and at the end of surgery in supine position, compared to the preoperative supine measurement (Figure 1). During the case, there was no significant difference in mean IOP between the 60, 90 and 120 minute measurements. Supine IOP measured after induction of anesthesia ranged from 9 to 28 mm Hg. IOP in steep trendelenburg at 120 minutes ranged from 25 to 54 mm Hg. Ocular perfusion pressure ranged from 50 to 84 mm Hg when supine at the start of surgery, and ranged between 21 and 63 mm Hg at 120 minutes. Average ocular perfusion pressure was lowest at 90 minutes (46.7 mm Hg) and 120 minutes (45.3 mm Hg). Ocular perfusion pressure returned to near baseline levels at the end of surgery (60.3 mm Hg vs. baseline of 62.3 mm Hg). In 17 cases that were > 3.5 hours long, IOP continued to rise (data not presented by author). Patients with a BMI > 30 kg/m2 had the highest ending IOP.



Figure 1. Comparison of Changes in MAP, IOP and OPP

Figure 1

Note. ST = steep trendelenburg. IOP = intraocular pressure, MAP = mean arterial blood pressure, OPP = ocular perfusion pressure. OPP = MAP – IOP.



In 26% of patients (n = 10) the IOP tripled within 2 hours in steep trendelenburg, with ocular perfusion pressure reportedly dropping below IOP at 90 minutes and not rising above IOP until the end of surgery. These patients had a BMI >30 kg/m2 (40.5% of sample had a BMI > 30 kg/m2). However, only 66.7% of obese patients in this study had their IOP triple (10 of 15 obese patients). Facial and periorbital edema was present in some patients, though this was not objectively measured. No patient experienced blindness. Differences in ocular perfusion pressure were not analyzed because MAP was maintained at around 80 mm Hg.


Conclusion This study provided a correlation between duration of surgery spent in steep trendelenburg and increased IOP. This lead to a reduced ocular perfusion pressure, which in some patients was below IOP, possibly increasing the risk of postoperative visual loss.



Annually, approximately 90,000 radical prostatectomies are performed, with 70,000 being done with robotic assistance (see Postoperative visual loss after robotic assisted laparoscopic procedures in steep trendelenburg position is an extremely rare, but devastating complication. In addition to this case, I could only find two other reported cases of postoperative visual loss, secondary to posterior ischemic optic neuropathy after minimally invasive prostatectomy performed in the steep trendelenburg position.1 Weber et al1 reported in these two cases the surgical duration was 6.5 and 9 hours. In the second case, the patient had an estimated blood loss of 1,200 mL and total fluids of 6,500 mL. The patient in the latter case had significant facial swelling after surgery. In the first case the patient lost vision posteriorly, but retained 20/20 vision in both eyes. The second patient had permanent postoperative visual loss. The authors postulated that the visual loss may have been related to the prolonged surgical duration in steep trendelenburg position and noted that impaired venous return has been implicated in ischemic optic neuropathy after radical neck dissection with jugular vein ligation. They commented that the steep learning curve with robotic assisted prostatectomies might be associated with these types of complications.


I commend the investigator of this study for describing changes in IOP during steep trendelenburg procedures. This is an important area that needs research. Fortunately, the incidence of postoperative visual loss is extremely rare. However, because the event is so rare it is difficult to investigate. This makes it challenging to develop evidence based practice guidelines for the prevention of this complication.


The findings of this study confirm the findings2 of increased IOP in steep trendelenburg position during robotic assisted laparoscopic prostatectomy (see Hemodynamic perturbations during robot-assisted laparoscopic radical prostatectomy in 45 degree trendelenburg position in Anesthesia Abstracts January 2011, Vol. 5, No. 1, 9-12).3 Awad et al2 reported that IOP on average was 13 mm Hg higher at the end of the procedure while still in steep trendelenburg (under GETA with abdominal insufflation) when compared to the supine position (patient awake). The duration of time in steep trendelenburg was 68 minutes (range 31-115 min).


Likewise, Molley in this study reported that all IOP measurements while in this position and prior to emergence from anesthesia were significantly higher than the initial baseline measurement. Awad et al2 reported that ETCO2 was the strongest predictor of increased IOP while in steep trendelenburg. Surgical duration was also a predictor, with IOP increasing 0.05 mm Hg for every minute of surgery. Increased ETCO2 is associated with choroidal vasodilation and elevated IOP. Steep trendelenburg position in combination with laparoscopy leads to increased CVP and most likely to increased orbital venous pressure. Recent research suggests that CVP doubles with insufflation of the abdomen and placement in steep trendelenburg position.3 Therefore, the rare cases of postoperative visual loss after these procedures may in fact be due to an “orbital compartment syndrome.”


A few limitations need to be pointed out in this study. The investigator included patients having multiple laparoscopic surgical procedures in steep trendelenburg position. While the pathophysiology is the same, these differences may have confounded the results. It is not clear from the design at what point the insufflation of the abdomen occurred. Also she reports the patients with BMI > 30 kg/m2 had the greatest increase in IOP and secondary decreases in ocular perfusion pressure. However, no demographic data is presented on this subgroup. I would have liked to know how they differed from other patients. There were significant changes over time in IOP, and IOP at all time points was higher than the initial measurement. However, there was no statistically significant increase in IOP after 60 minutes. The investigator however, concluded that there was a correlation between the duration of surgery spent in steep trendelenburg and increased IOP. The investigator did not run correlations, only looked at changes over time, which is not the same thing. Therefore, I disagree with this conclusion. Although, previous research does suggest a relationship exists between surgical duration and IOP.2


Despite these limitations, this study does add to our understanding of the changes seen with IOP during steep trendelenburg procedures.

Dennis Spence, PhD, CRNA

1. Weber ED, Colyer MH, Lesser RL, Subramanian PS. Posterior ischemic optic neuropathy after minimally invasive prostatectomy. J Neuro-Opthalmol 2007;27:285-287.

2. Awad H et al. The effects of steep trendelenburg positioning on intraocular pressure during robotic radical prostatectomy. Anesth Analg 2009;109:473-8.

3. Lestar M, Gummarsson L, Lagerstrand L, Wiklund P, Odeberg-Wernerman S. Hemodynamic perturbations during robot-assisted laparoscopic radical prostatectomy in 45º trendelenburg position. Anesth Analg Published ahead of print January 2011.

Editor’s Note: Previous research has shown that end tidal CO2 does not accurately represent arterial CO2 during prolonged laparoscopic surgery in Trendelenburg position. The disconnect between arterial CO2 and end tidal CO2 becomes greater the longer the duration of the case. For more details see, “Laparoscopic colon surgery :unreliability of end-tidal CO2 monitoring,” in the August 2008 issue of Anesthesia Abstracts.


The views expressed in this article are those of the author and do not reflect official policy or position of the Department of the Navy, the Department of Defense, the Uniformed Services University of the Health Sciences,  or the United States Government.

© Copyright 2011 Anesthesia Abstracts · Volume 5 Number 5, May 31, 2011

Obstetric Anesthesia
Anesthesia-related maternal mortality in the united states: 1979-2002

Obstet Gynecol 2011:117:69-74

Hawkins JL, Chang J, Palmer SK, Gibbs CP, Callaghan WM


Purpose The purpose of this study was to compare anesthesia-related maternal mortality from 1979-1990 with data from 1991 to 2002; and to compare the case-fatality rates of general and regional anesthesia during cesarean delivery.


Background Anesthesia-related causes of maternal mortality have progressively declined and now account for less than 2% of all pregnancy-related deaths in the United States. Reviews of anesthesia-related maternal mortality, advances in obstetric anesthesia, and incorporation of team training and simulation on labor and delivery have improved patient safety. However, adverse events still occur. This study sought to examine trends in anesthesia-related maternal mortality over time and to compare differences in mortality rates associated with general and regional anesthesia during cesarean delivery.


Methodology This was a retrospective review of maternal deaths related to anesthesia from 1991 to 2002. Investigators used the Pregnancy Mortality Surveillance System established by the Centers for Disease Control and Prevention (CDC) Division of Reproductive Health to identify maternal deaths caused by anesthesia-related complications. From 1991-2002 there were 5,946 pregnancy-related deaths, with 86 determined to be caused by anesthesia-related complications. Three obstetrical anesthesiologists independently reviewed each death certificate of these 86 anesthesia-related cases to confirm the cause of death. After the initial review, 27 early losses were excluded, leaving 56 cases associated with an obstetric delivery (live birth or stillbirth).


Pregnancy-related mortality rates per million live births were calculated using national data on live births from the 1991-2002 natality files of the National Center for Health Statistics. Case fatality rates for general and regional anesthesia for cesarean delivery were estimated. Pregnancy-related deaths associated with anesthesia during obstetric deliveries for 1991-2002 were compared to 1979-1990.


Result Between 1991 and 2002, 1.6% (n = 86) of all pregnancy-related deaths in the United States were related to complications associated with anesthesia. The overall rate of deaths due to anesthetic complications decreased by 59%; from 2.9 per million live births for 1979-1990 to 1.2 per million live births for 1991-2002 (Figure 1). Comparison of demographic and pregnancy characteristics between the two time periods indicated that maternal race and prenatal care were comparable. For 1991-2002 there were more maternal deaths in the 30-34 year old and 35-39 year old groups compared to the 1979-1990 data (Table 1). Nearly all patients who died from anesthesia-related complications during 1991-2002 were undergoing cesarean delivery (86%). This rate was similar to the 1979-1990 rate (82%).




Figure 1. Pregnancy-Related Mortality Ratio due to Anesthesia

Figure 1




Table 1. Demographic and Pregnancy Characteristics for Pregnancy-Related Deaths Associated with Anesthesia-Related Complications



(n = 129)


(n = 56)

Age (years)








16 (12.4%)

41 (13.8%)

36 (27.9%)

25 (19.4%)

6 (4.7%)

5 (3.9%)


7 (12.5%)

15 (26.8%)

9 (16.1%)

13 (23.2%)

10 (17.9%)

2 (3.6%)



African American




58 (45%)

67 (51.9%)

4 (3.1%)



24 (42.9%)

26 (46.4%)

2 (3.6%)

4 (7.1%)

Delivery Procedure

Cesarean delivery

Vaginal delivery



106 (82.2%)

6 (4.7%)

17 (13.2%)


48 (85.7%)


8 (14.3%)*

Note. *Delivery procedures not determined, however no known deaths were associated with vaginal delivery.



Anesthesia-related deaths during cesarean delivery were consistently higher when general anesthesia was administered compared to regional anesthesia (Figure 2). However, case fatality for general anesthesia decreased from 20 deaths per million general anesthetics during 1979-1984 to 6.5 deaths per million general anesthetics during 1997-2002. Similarly, there was a decrease from 8.6 deaths per million regional anesthetics during 1979-1984 to 3.8 deaths per million regional anesthetics during 1997-2002. There was a slight increase in regional anesthesia case fatality rates between the years 1991-1996 and 1997-2002 (Figure 2). The relative risk of death due to general anesthesia compared to that for regional anesthesia decreased from 6.7 (95% CI 3-14.9) for 1991-1996 to 1.7 (95% CI 0.6-4.6) for 1997-2002.




Figure 2. Case Fatality Rates of Anesthesia-Related Deaths during Cesarean Deliveries by Type of Anesthesia

Figure 2

Note. Deaths per million general or regional anesthetics.



Overall, the leading causes of death for 1991-2002 were intubation failures or induction problems (23%), followed by respiratory failure (20%), and high spinal or epidural block (16%). However, when causes were differentiated by type of anesthesia administered, approximately 67% of deaths associated with general anesthesia were caused by intubation failure or induction problems. In contrast, in women who died after regional anesthesia for cesarean delivery, 26% of deaths were caused by high spinal or epidural, 19% by respiratory failure, and 19% by a drug reaction.



Conclusion Maternal deaths related to anesthesia complications decreased by almost 60% between the periods 1979-1990 and 1991-2002. Case fatality rates for general anesthesia for cesarean delivery decreased, though death rates for regional anesthesia increased.



This study demonstrated that obstetric mortality related to anesthesia causes is now less than 1.7% for all deliveries, and case fatality rates for general and regional anesthesia for cesarean delivery have also dramatically decreased. The reason for these decreased mortality rates are related to the increased use of neuraxial anesthesia and decreased need for general anesthesia for cesarean delivery. Additionally, increased awareness of predictors of difficult airway, advances in airway equipment (i.e., LMA), and approach to the difficult airway have helped decrease maternal mortality related to general anesthesia.


Despite these increases, the leading causes of mortality continue to be due to intubation failure or induction problems. It is important to point out that this time frame was between 1991-2002, and since then newer neuraxial anesthetic techniques (i.e., combined spinal epidural) and airway equipment (i.e., indirect video laryngoscopy, Glidescope) have become more widespread. I suspect that with the next analysis the mortality rate for general anesthesia will be lower. General anesthesia is typically reserved for emergent need for cesarean delivery. Sometimes there is limited time to evaluate the patient. Typically, the patient does not have an epidural or there is not enough time to attempt to administer a spinal. Additionally, there may be contraindications to neuraxial anesthesia (i.e., hemorrhaging parturient). Furthermore, because of the increased use of epidural anesthesia, anesthesia trainees are graduating with little or no experience administering general anesthesia for cesarean delivery. These factors combined contribute to the higher mortality rate when compared to neuraxial anesthesia for cesarean delivery.


So what can the anesthesia provider do to help further reduce the mortality rate associated with anesthesia in the pregnant patient? The first step is to conduct a thorough preoperative evaluation, with special attention to identifying patients with a potential difficult airway. Second, when administering epidural anesthesia, providers should administer incremental boluses (i.e., 5 mL at a time), and if a high or total spinal occurs the symptomatic bradycardia should be aggressively treated. Anesthesia providers need to be confident that if a patient has a labor epidural it will work for a cesarean delivery. A saying I tell my students is “when in doubt, pull it out and replace it.” Difficult airway equipment should be immediately available, preferably in the operating room, whenever a cesarean delivery is required. Given recent research that suggests devices such as the Glidescope result in higher intubation success, anesthesia providers may want to consider obtaining one for their labor suites. Finally, it is essential to maintain close communication with the nurses, midwives, and obstetricians and ensure a proper patient hand-off occurs between anesthesia providers.

Dennis Spence, PhD, CRNA

The views expressed in this article are those of the author and do not reflect official policy or position of the Department of the Navy, the Department of Defense, the Uniformed Services University of the Health Sciences,  or the United States Government.

© Copyright 2011 Anesthesia Abstracts · Volume 5 Number 5, May 31, 2011

Pediatric Anesthesia
Polysomnographic variables predictive of adverse respiratory events after pediatric adenotonsillectomy

Arch Otolaryngol Head Neck Surg 2011;137:15-18

Jaryszak EM, Shah RK, Vanison CC, Lander L, Choi SS


Purpose The purpose of this study was to identify predictors of adverse respiratory events after pediatric adenotonsillectomy in patients who had a preoperative polysomnography (PSG) examination.


Background A majority of pediatric patients who have adenotonsillectomy surgery (T&A) can be discharged home on the same day. However, there is a subset of patients who may need extended observation because of obstructive sleep apnea syndrome (OSAS) and/or other comorbidities. Common respiratory difficulties range from desaturation due to swelling and secretions to postobstructive pulmonary edema. There is evidence to suggest that younger patients, those with significant co-morbidities, patients with high apnea-hypopnea indexes during a sleep study (AHI), and lower nadir oxygen saturations are at increased risk for respiratory complications.


Methodology This was a retrospective case-control study of pediatric patients who underwent preoperative polysomnography (sleep study) prior to T&A. The study was conducted at Children’s National Medical Center in Washington, DC. A total of 1,131 records of pediatric T&A procedures performed by two ENT surgeons were reviewed. Of those, 151 pediatric patients underwent a preoperative sleep study. No rigid criteria were used to select patients for preoperative polysomnography. Patients not uncommonly presented to the ENT clinic with sleep study results.  At this institution, patients were admitted for overnight observation following a T&A if they were less than 5 years old, had significant comorbidities, or had significant OSAS.


A retrospective review of these 151 patients was completed to identify patients with postoperative respiratory complications. Respiratory complications were defined as:

  1. oxygen desaturation requiring supplemental oxygen therapy
  2. oxygen desaturation requiring medication
  3. oxygen desaturation requiring nonrebreather or positive pressure (CPAP or BiPAP)
  4. oxygen desaturation requiring intubation
  5. postobstructive pulmonary edema.


Logistic regression was performed to determine factors predictive of the presence or absence of one of these respiratory complications. A P < 0.05 was considered significant.


Result Of the 1,131 pediatric patients who had a T&A between July 2006 and December 2008, 13.4% (n=151) underwent preoperative polysomnography. Medical comorbidities in these 151 patients included asthma (52), neurological disorder (19), reflux disease (14), Trisomy 21 (13), cardiac disorder (8), psychiatric disorder (8), prematurity (8), other syndromes (4), and other diagnoses (10). (The investigators did not specify whether or not patients had more than one of these comorbidities). Of the 151 patients who underwent polysomnography, 15.2% (n=23) experienced adverse respiratory events; with some experiencing more than one complication. The most common respiratory complication was oxygen desaturation requiring supplemental oxygen therapy (n = 21; 13.9%), oxygen desaturation requiring medication (n = 1; 0.7%), oxygen desaturation requiring CPAP or BiPAP (n = 2; 1.3%), and postobstructive pulmonary edema (n = 1; 0.7%). No patient experienced oxygen desaturation requiring intubation.


The 23 patients in the respiratory complication group had significantly higher AHI scores (31.8 vs. 14.1, P = 0.001), higher hypopnea index (22.6 vs. 8.9, P = 0.004), lower nadir oxygen saturations (72% vs. 84%, P = 0.001), and a higher BMI greater than the 95th percentile (48% vs. 30%, P = 0.02) compared to the 128 patients who did not experience a respiratory complication.  No difference in age was found between the two groups (mean age 5.8 years in both groups, P = 0.98). Patients who experienced a respiratory complication were 32.1 times more likely to require a prolonged hospital course (95% CI: 7.8-131.4, P <0.05). Patients in the respiratory complications group spent a total of 22 additional days in the hospital beyond routine overnight observation. Patients without a respiratory complication spent a total of only 4 additional days in the hospital beyond routine overnight observation; 3 days for poor oral intake and 1 night for planned admission for a 13-month-old patient (1 ± 0.3 days vs. 0.03 ± 0.02 days). While most respiratory complications were oxygen desaturation requiring supplemental oxygen, 5.9% of patients (9 of 151) required admission and close monitoring.


Conclusion Polysomnography may help identify pediatric patients at higher risk for adverse respiratory events. This information may be useful in planning anesthetic and postoperative management. Predictors of respiratory complications included a higher AHI, hypopnea index, or BMI, and a lower nadir oxygen saturation.



These results demonstrated that pediatric patients with severe OSAS (AHI >30) were at increased risk for postoperative respiratory events after T&A. This is not surprising given what is known about OSAS. One might expect that once the hypertrophied tonsillar tissue was removed that obstructive symptoms may be less; however tissue trauma and resultant edema may still contribute to airway obstruction and adverse respiratory events. As in adults after uvulopharnygoplasty surgery, it may take weeks or months for OSAS symptoms to resolve after T&A. Higher BMI may also increase the risk.


While I think the study results are important, there are several limitations. There probably was some selection bias as many of these patients received polysomnography prior to surgery. It may be that the patient’s pediatricians or ENT surgeon wanted additional information documenting OSAS prior to performing the T&A. Looking at the demographics one can see that many of the patients had significant comorbidities which may have either prompted the polysomnography or contributed to the outcomes. Thus the sample may not be representative of the general population. It would have been nice if the investigators had presented the ASA status of the patients. One question I have, was the ASA status higher in the group of patients who had adverse events? Other information that is lacking from the study was the amount of opioid administered during the perioperative period, the airway exam or degree of preoperative tonsillar hypertrophy, and timing of the adverse events. Most of the complications were oxygen desaturation requiring supplemental oxygen therapy. Did these events occur early in the PACU or late on the ward? It is important to point out that one patient experienced postobstructive pulmonary edema, which is a rare complication that can occur after relief of severe obstruction. It would have been helpful if the investigators had provided more information about the patient that experienced this complication.


Finally, I think the investigators analysis plan was not consistent with the results they presented. They proposed to perform a logistic regression to determine which variables (e.g., AHI) predicted whether or not a patient experienced an adverse respiratory event. However, the investigators only presented group differences. I assume they used a t-test or nonparametric equivalent test to analyze these results. It would have been more helpful if the investigators had presented the odds ratio for the variables of interests (i.e., AHI, BMI, number of comorbidities) as this would enable the reader to understand the magnitude of the risk and better apply the results to clinical practice.


At my institution it is rare that a child receives a preoperative polysomnograph. If a child is less than 5 years typically they are having a T&A because of OSAS symptoms rather than for tonsillitis. It is essential during the preoperative evaluation to ask the surgeon about the reason for the T&A and to question the parents regarding the severity of their child’s OSAS symptoms. If OSAS symptoms are present, then I limit the amount of opioid to a maximum of 1 µg/kg of fentanyl for the case and consider prolonged monitoring depending on the severity of symptoms and the comorbidities. At my institution most children less than 5 years old are admitted overnight for airway observation because we assume the reason for surgery is OSAS. Anesthesia providers should evaluate their facilities policies regarding the level of monitoring required after airway surgery in pediatric patients with OSAS.

Dennis Spence, PhD, CRNA

The views expressed in this article are those of the author and do not reflect official policy or position of the Department of the Navy, the Department of Defense, the Uniformed Services University of the Health Sciences, or the United States Government.

© Copyright 2011 Anesthesia Abstracts · Volume 5 Number 5, May 31, 2011

Low-dose ketamine with multimodal postcesarean delivery analgesia: a randomized controlled trial

Int J Obst Anesth 2011;20:3-9

Bauchat JR, Higgins N, Wojciechowski KG, McCarthy RJ, Toledo P, Wong CA


Purpose The purpose of this study was to determine if a single low-dose of intravenous ketamine 10 mg administered during cesarean delivery would decrease the incidence of breakthrough pain in the first 24 hours postoperatively.


Background Pain after cesarean delivery can delay recovery and interfere with a mothers ability to care for her baby. Multimodal analgesic techniques which include intrathecal and systemic opioids and nonsteriodal anti-inflammatory drugs (NSAIDS) are usually administered to decrease postoperative pain while minimizing side effects from large dosages of single agents. Ketamine in sub-hypnotic dosages has been used to decrease acute and chronic pain. Multiple investigations have demonstrated that ketamine decreases opioid requirements and pain scores in the first 24 hours after surgery. There are limited studies on the efficacy of sub-hypnotic doses of ketamine administered in a multimodal analgesic regimen after cesarean delivery.


Methodology This was a randomized, double-blind, placebo controlled trial of 188 ASA I-II parturients at ≥ 37 weeks gestation scheduled for elective cesarean delivery under spinal anesthesia with intrathecal morphine. Parturients were excluded if they had a history of chronic pain, substance abuse, chronic opioid use, history of substance abuse, or BMI ≥ 40 kg/m2. Parturients were randomized to receive intravenous ketamine 10 mg or placebo (normal saline). Study medications were prepared by a blinded anesthesia provider and administered in an infusion pump over 10 minutes.


A blinded anesthesia provider administered the spinal anesthetic at the L3-4 interspace. The dose was 12 mg of 0.75% hyperbaric bupivacaine with 150 µg of preservative free morphine. Five minutes after delivery of the infant, study medications were administered. Postoperatively all parturients were administered ketorolac 30 mg every 6 hours for 24 hours (could refuse) and then ibuprofen 600 mg every 6 hours as needed between 24 and 72 hours. For breakthrough pain parturients could request 1 tablet of acetaminophen 325 mg/hydrocodone 10 mg every 4 hours prn. If pain relief was ineffective, parturients could request an additional tablet 1 hour later.


Five minutes after administration of the study drugs, subjects were evaluated for the presence of nausea, vomiting, pruritus, psychomimetic symptoms, and sedation using the Richmond agitation-sedation scale. Breakthrough pain was defined as need for supplemental pain medication. The lysergic acid diethylamine short form of the Addiction Research Center Inventory (LSD-ARCI) was administered at 1 and 4 hours postoperatively to evaluate subjective psychomimetic effects. Data collected included pain scores, time to first analgesic request, satisfaction with analgesia at 24 and 72 hours, and cumulative number of acetaminophen/hydrocodone tablets administered at 24, 48 and 72 hours. Patients were contacted at 2 weeks to evaluate satisfaction with analgesia and average pain score. Sample size and statistical analysis were appropriate. A P < 0.05 was considered significant.


Result A total of 177 parturient results were analyzed, n = 85 in the ketamine group and n = 89 in the placebo group. No significant differences were found in baseline demographics. Median age was 34 (31-37), and BMI = 77 kg/m2 (71-87). Approximately 67% had a prior cesarean delivery. No significant differences were found in the incidence of nausea, vomiting, or pruritus between the two groups (P = NS). Richmond agitation-sedation scores were significantly different between the two groups (P = 0.003; Table 1).




Table 1. Richmond Agitation-Sedation Scale Results



n = 85


n = 89





P = 0.003







Spontaneous complaints



P < 0.001

Disturbing dreams



P = NS

Note. 1 = anxious or apprehensive but movements not aggressive or vigorous, 0 = alert and calm, -1 = not fully alert, but has sustained awakening with eye contact, to voice. Spontaneous complaints included: lightheadedness, double vision, dizziness.



For the primary outcome, incidence of breakthrough pain within the first 24 hours, 75% in the ketamine group and 74% in the placebo group experienced breakthrough pain (P = NS). Median time to first analgesic request was 684 minutes in the ketamine group (95% CI: 337-1031) and 760 minutes in the placebo group (95% CI: 346-1174) (P = NS). Pain scores at first analgesic request were similar between the two groups, with median scores of 3 in the ketamine group (2-4) and 4 in the saline group (2-5) (P = NS). Cumulative acetaminophen/hydrocodone tablets consumed at 24, 48 and 72 h were similar (Figure 1; P = NS). Pain scores were similar between the two groups at all time points in the first 24 hours (P = NS; Figure 2). Pain scores at 2 weeks were significantly lower in the ketamine group (median difference: -0.6, 95% CI: -1.1 to -0.09, P = 0.02). Satisfaction scores were similar at all time points (P = NS).




Figure 1. Comparison of Pain Scores

Figure 1


Figure 2. Cumulative Pain Tablets

Figure 2



Conclusion A single low intravenous dose of ketamine 10 mg was not efficacious in decreasing pain or opioid consumption when compared to placebo after spinal anesthesia in parturients receiving a multimodal analgesic regimen. Pain scores were lower at 2 weeks postpartum in the ketamine group, suggesting a ketamine effect on the development of chronic pain after cesarean delivery.




Multimodal analgesic regimens with intrathecal morphine, NSAIDS, and Percocet tablets are highly effective in reducing pain after cesarean delivery. This study sought to determine if adding a sub-hypnotic dose of ketamine would further decrease pain and opioid consumption beyond what a multimodal regimen can provide. The investigators found that the addition of a single 10 mg dose after delivery of the neonate had no effect on any of their primary or secondary outcomes. Parturients did, however, experience more psychomimetic symptoms shortly after administration when compared to placebo. The investigators did find a -0.6 point lower pain score in the ketamine group at 2 weeks. While this is statistically significant, I do not consider such a small effect to be clinically significant. Based on these results I would not include ketamine as part of a multimodal regimen.


Some investigators have found sub-hypnotic doses of ketamine might decrease the incidence of chronic pain. The mechanism is thought to be through decreased central sensitization. Given cesarean delivery is associated with 6-8% incidence of chronic pain, future studies are clearly needed. Additionally, future studies should evaluate a weight based sub-hypnotic dose of ketamine (e.g., 0.15 mg/kg).

Dennis Spence, PhD, CRNA

The views expressed in this article are those of the author and do not reflect official policy or position of the Department of the Navy, the Department of Defense, the Uniformed Services University of the Health Sciences, or the United States Government.

© Copyright 2011 Anesthesia Abstracts · Volume 5 Number 5, May 31, 2011

Rolapitant for the prevention of postoperative nausea and vomiting: a prospective, double-blinded, placebo-controlled randomized trial

Anesth Analg 2011;112:804-12

Gan TJ, Gu J, Singla N, Chung F, Pearman MH, Bergese SD, Habib AS, Candiotti KA, Mo Y, Huyck S, Creed MR, Cantillon M, Rolapitant Investigation group


Purpose The purpose of this Phase II dose-finding study was to evaluate what effect Rolapitant had on the incidence of postoperative nausea and vomiting (PONV) in adult women undergoing elective open abdominal surgery.


Background Rolapitant is a potent, selective neurokinin-1 (NK-1) receptor antagonist with a half-life of 180 hours. It has minimal drug interactions and is well tolerated in dosages up to 200 mg in healthy volunteers. Similar to aprepitant, rolapitant blocks the binding of substance P at NK-1 receptors. Substance P has been found to stimulate the vomiting reflex in the nucleus tractus solitarii and area postrema. This Phase II study sought to determine the optimal dosage, efficacy, safety and tolerability of rolapitant when used for up to 5 days.


Methodology This was a prospective, randomized, double-blind, placebo-controlled trial of 619 ASA I to III women at high risk for PONV scheduled for open abdominal surgery. Patients enrolled in the study were expected to be admitted to the hospital for at least 24 hours, and to require postoperative intravenous patient controlled analgesia. Patients were randomized to one of six study arms:

  1. rolapitant 5 mg (n = 103)
  2. rolapitant 20 mg (n = 102)
  3. rolapitant 70 mg (n = 103)
  4. rolapitant 200 mg (n = 104)
  5. ondansetron 4 mg (n = 104)
  6. placebo (n = 103)


Rolapitant was administered no later than 30 minutes before induction of anesthesia and ondansetron administered intravenously immediately before induction of anesthesia. Anesthesia was induced with propofol and neuromuscular blocker (provider choice), and maintained with inhalation agent and up to 50% nitrous oxide and opioid as needed. Postoperatively, patients received intravenous PCA with morphine, or if allergic to morphine; fentanyl, hydromorphone, or demerol. Rescue antiemetic was ondansetron; it was administered when nausea severity was ≥ 4.


The primary outcome of this study was the incidence of no emetic episodes for 24 hours after extubation. Secondary outcomes included the response rate for no emetic episodes for up to 120 hours, incidence of complete response (defined as no emesis or need for antiemetic rescue medication), incidence of no nausea for up to 120 hours after extubation, time to first rescue medication, time to first emetic event, and time to first significant nausea (≥ 4 out of 10). Safety and tolerability were also assessed. Statistical analysis was appropriate. An intention to treat and per protocol analysis was conducted. A P < 0.05 was significant.


Result Result A total of 532 patients were included in the analysis. No significant differences were found in demographics or baseline characteristics between the 6 groups. For the primary outcome, patients who received rolapitant 20 mg, 70 mg, or 200 mg had a significantly lower incidence of emesis between 0 and 24 hours after surgery when compared to placebo (P < 0.05; Figure 1). The 70 mg and 200 mg groups demonstrated a significantly lower incidence of emesis for up to 120 hours (P < 0.05). When compared to placebo, patients taking the 70 mg dose were 2.87 times less likely to experience an emetic episode in the first 24 hours (95% CI: 1.54-5.25, P < 0.001). For the 200 mg dose the odds ratio for no emesis was 4.73 (95% CI: 2.37-9.42, P < 0.001).




Figure 1. Incidence of No Emesis

Figure 1



Patients who received rolapitant 70 mg or 200 mg had a higher incidence of complete response (absence of emesis or use of antiemetic rescue) between 72 and 120 hours after extubation compared to placebo (P < 0.05; Figure 2). Ondansetron was similar to placebo between 72 and 120 hours (P = ns). At 48 hours, the 200 mg rolapitant group had the highest complete response rate, with 37% (n = 38) experiencing no emesis or need for rescue antiemetic (P < 0.05). Overall, patients who received rolapitant (all doses except 70 mg) had a longer time to first emetic episode compared to placebo (P < 0.001). When compared to placebo, patients who received rolapitant 200 mg had the longest time to first emetic episode (28.3 ± 33.5 hrs vs., 14.9 ± 22 hrs, P < 0.001).




Figure 2. Incidence of Complete Response

Figure 2

Note. Patients who received rolapitant 70 mg had a significantly higher incidence of complete response (no vomiting or need for rescue antiemetic) between 0 and 72 hours when compared to placebo (P < 0.05). Patients who received rolapitant 120 mg had a significantly higher incidence of complete response at all time periods except between 0 and 24 hours when compared to placebo (P < 0.05).



The incidence of nausea was similar between the groups at all time points (P = ns). The incidence of PONV between 0 and 24 hours was similar between the groups (P = ns). Patients in the 70 mg rolapitant group had a significantly lower incidence of nausea at 120 hours and no PONV between 0 and 24 hours when compared to placebo (P < 0.05).


No significant differences were found in the incidence of side effects or adverse events between the groups (P = ns). No ECG or hemodynamic changes were found with any of the rolapitant groups when compared to placebo (P = ns).


Conclusion Rolapitant decreased the incidence of postoperative vomiting and emetic episodes in a dose-dependent manner when compared to placebo. No adverse events or safety concerns were found with doses up to 200 mg. The optimal rolapitant dose to prevent PONV is not yet known.




Rolapitant is a new NK-1 receptor antagonist which has prolonged antiemetic effects; up to 120 hours. It is similar to aprepitant; however it has a longer duration of action, 180 hours vs. 72 hours, which makes it useful for preventing postoperative emesis for an extended period of time. This could be useful in patients in which vomiting could result in significant morbidity (i.e., craniotomy, inner ear surgery). Similar to aprepitant, rolapitant appears to have minimal efficacy in decreasing the incidence of postoperative nausea when compared to placebo.


This study was a Phase II dose finding study examining a range of rolapitant doses in high risk patients undergoing open abdominal surgery. The investigators found that the largest doses, 70 and 120 mg, had the greatest effect on decreasing the incidence of vomiting and emesis. It is important to point out that rolapitant is still under clinical investigation and is not yet approved by the FDA. Further research is needed to identify the optimal antiemetic dose. Hopefully in the next few years this drug will come to market. Unfortunately, it will probably be just as expensive as aprepitant, which costs approximately $102 for an 80 mg capsule. However, if I just had surgery I would be willing to pay this amount to keep from vomiting.

Dennis Spence, PhD, CRNA

There are 4 phases of clinical trials. Clinical trials of investigational new drugs are conducted to establish the efficacy and safety of the drug. Below are descriptions of the phases taken from the FDA website


  • Phase I trials test an experimental drug in a small group of people (20-80) for the first time to evaluate its safety, dosage range, and identify side effects.
  • In Phase II trials researchers give the study drug to a larger group of people (100-300) to see if it is effective and to further evaluate its safety.
  • In Phase III trials the study drug is given to large groups of people (1,000 - 3,000) to confirm its effectiveness, side effects, compare it to commonly used treatments, and collect information that will allow the experimental drug to be used safely.
  • In Phase IV trials, post marketing studies delineate additional information including the drug's risks, benefits, and optimal use.


The views expressed in this article are those of the author and do not reflect official policy or position of the Department of the Navy, the Department of Defense, the Uniformed Services University of the Health Sciences, or the United States Government.

© Copyright 2011 Anesthesia Abstracts · Volume 5 Number 5, May 31, 2011

Dreaming in sedation during spinal anesthesia: A comparison of propofol and midazolam infusion

Anesth Analg. 2011;112:1076-1081

Kim DK, Joo Y, Sung TY, Kim SY, Shin HY


Purpose The purpose of this study was to compare propofol and midazolam sedation during spinal anesthesia and determine their impact on dreaming and patient satisfaction.


Background Dreams can be created by internal and external stimuli and are often expressions of emotional preoccupation. Dreams that are reported after sedation and general anesthesia often have similar features to those experienced in natural sleep. However, studies of dreaming during general anesthesia indicate that patients find them unpleasant and they sometimes lead to postoperative depression. It is possible that the risk of unpleasant dreams during neuraxial anesthesia with sedation may negate the positive benefits of providing sedation, in some cases. This study examined the impact that propofol and midazolam had on dreaming and patient satisfaction when used with neuraxial anesthesia.


Methodology Patients were selected for the study who were scheduled for any procedure under spinal anesthesia. Their ages ranged from 18 to 70 years old, and their physical status ranged from I to III. Patients with cognitive deficits, intellectual impairment, and those who had taken hypnotics were excluded. Two hundred twenty patients were enrolled in the study. Patients were randomly assigned sedation with either propofol or midazolam. The propofol group was given an infusion of 10 mg/kg/hr until their eyes closed, and then the infusion was reduced to 5 mg/kg/hr until the end of surgery. Patients assigned midazolam were given an infusion of 1.0 mg/kg/hr which was reduced to 0.5 mg/kg/hr after their eyes closed, and continued until the end of surgery. Adjustments to the maintenance infusion of both drugs were based on the following widely accepted sedation scale ranging from 1 to 5, where:

1 = fully awake

2 = drowsy

3 = eyes closed and responsive to verbal stimulation

4 = eyes closed but responsive to mild stimulation

5 = unresponsive to mild stimulation.


The infusions were cut in half when the score reached 4 to 5, and doubled when the score was 1 to 2.

After the procedure and when patients were oriented, they were asked a set of standard questions:

1. What was the last thing you remember before going to sleep?

2. What was the first thing you remember when you woke up?

3. Can you recall anything in between?

4. Did you have any dreams during sedation?


The patients were asked the same questions 30 minutes later. The narratives from these questions were collected and evaluated for association between narrative content and surgical events. Patients who had dreams were asked to rate the content of those dreams in relationship to emotional content, memorability, visual vividness, emotional intensity, and strangeness. Finally, patients were asked to rate their satisfaction with sedation on a scale from 0 to 100 with 0 being no satisfaction and 100 being fully satisfied.


Result Five patients were excluded because of conversion to general anesthesia which left 108 subjects in the propofol group and 107 subjects in the midazolam group. Patient characteristics, surgical length, and sedation requirements were similar in both groups. There were 43 dreamers (40%) in the propofol group and 13 dreamers (12%) in the midazolam group (p<0.001). Only 6 patients out of both groups reported dreaming at the second interview 30 minutes after emergence. Simple and pleasant dreams were reported by 64% of patients who dreamed; dreams of being pursued were reported by 11%, dreams associated with an operative procedure were reported by 9%, and 5% could not remember the details of the dream. Those receiving propofol had more memorable and vivid dreams. Patient satisfaction was high and similar in both groups.


Conclusion Previous studies with propofol had reported a dream rate of 35-36%. The reported rate of 40% in this study was similar. In this study, midazolam had a significantly reduced incidence of dreaming; an absolute reduction of 28% compared to propofol (approximately 12% vs 40%). The difference between the two drugs might be associated with the more rapid awakening for those administered propofol, which might have allowed them to more clearly recall their dreaming experience.

The authors of this study hypothesized that dreams during sedation might be unpleasant, but they found that most of the dreams were simple and pleasant (64%). Contrary to reports that suggest intraoperative dreaming can be confused with awareness, no such findings were reported in this study, and any dream containing components of a surgical procedure were found to be unrelated to the surgical events. Although there was a significant difference between the two groups related to dreaming, both groups had a high rate of satisfaction with the sedation. This finding was different than studies concerning dreaming and general anesthesia where patients confused dreaming with awareness. Because the results of this study indicated dreaming did not interfere with positive patient satisfaction, the authors suggest that providers should not be concerned about the possibility of dreaming when considering sedation during spinal anesthesia.



I was initially surprised by what I considered to be massive amounts of sedative given to these patients. The propofol dose started with 11 mg per minute in an average adult, or as high as 700 mg per hour. I use 200 mg vials and seldom use more than one vial per procedure. They would have used 3.5 vials per hour in the highest doses. The midazolam dose was even more concerning. Giving 70 mg of midazolam to an average adult in an hour would be considered a very large dose. Even at the lowest doses, a patient would receive no less than 20 mg of midazolam for an hour procedure. I am surprised these patients even remembered their own name after such large doses. I did not consider these doses to be standard, and there was no apparent attempt to compare drug doses to other published studies.


Although this study concluded that their patient population reported a high degree of satisfaction, they reported the rate of simple and pleasant dreams to be about 65%. I do not consider 35% of the study group having other than simple and pleasant dreams to be an insignificant amount. They downplayed the impact that dreams about a surgical procedure or dreams about being pursued may have had on the psychological wellbeing of a patient. The bottom line is that dreaming does have an impact on patients, and we do not fully understand what that impact may be, or how to prevent it.


One of the most painful events in my career occurred over a case of dreaming during sedation. I provided propofol sedation to a claustrophobic patient undergoing a 2 hour enclosed MRI. She was monitored very carefully, and I frequently entered the room to visually check her status. She appeared comfortable throughout the procedure, and at the end of the procedure she awakened without any complaint. Later that day, she was telling members of her family, other health care providers, and people in the community that she was screaming for help during the procedure. Although it could not be further from the truth, she believed I had ignored her. She was very angry about the event. I met with her, the radiology tech, and the patient advocate at our hospital to explore the problem. She would not believe our account of the situation and threatened to sue me and the hospital. I apologized for her perception of the events and tried to explain to her that stress over the procedure probably triggered post procedure dreams that felt very real to her. She never forgave me, and this simple situation became a very low point in my career. I have since changed my dosing strategies for sedation to include more midazolam and less propofol. I also council all patients about the possibility of dreaming and the reason dreams might occur.


This anecdote illustrates the risks of recall and dreaming in some patient situations. This study, and previous work, shows that patients “sedated” with propofol are more likely to have “memorable and vivid dreams” than those sedated with midazolam. Furthermore, the dreams need not be rooted in reality. The fact that in this study most dreams were observed to be “pleasant” does not indicate that “dreaming” during sedation will usually be so. Psychologists tell us that dreams can be influenced by sensory and cognitive stimuli immediately before or during sleep / sedation. While the investigators opined that we need not be concerned with the chance of dreaming during sedation, I disagree. I question our ability to accurately predict the likelihood of disturbing dreams. As a result, I think we would be better to employ sedation techniques that minimize dreaming and recall. We should keep in mind that while propofol may prevent recall in some patients, it is not as dependable an amnestic as benzodiazepines, ketamine, or potent inhalation agents. In general, I recommend that we use sedation techniques that minimize the probability of dreaming, especially in patients at risk for unpleasant dreams.


Most studies about dreaming and anesthetic agents indicate that patients receiving propofol alone have a higher rate of dreaming. We need to consider that fact when choosing our agents, and the risk of a patient having a bad dream during a surgical or diagnostic procedure.

Steven R Wooden, DNP, CRNA

© Copyright 2011 Anesthesia Abstracts · Volume 5 Number 5, May 31, 2011

Regional Anesthesia
Uncomplicated removal of epidural catheters in 4365 patients with international normalized ratio greater than 1.4 during initiation of warfarin therapy

Reg Anesth Pain Med 2011;36:231-235

Lui SS, Bukanendran A, Visusi ER et al


Purpose The purpose of this large, multi-site, observational study was to describe the incidence of spinal hematoma in patients on warfarin with an INR >1.4 at the time of epidural catheter removal.


Background Between 437,000 and 680,000 total joint replacement surgeries are performed annually in the United States. Epidural anesthesia is a common form of analgesia after total joint replacement. However, because these patients are at risk for thromboembolism, anticoagulation is required for prophylaxis. Warfarin is commonly initiated the night after surgery or on the first postoperative day. Current American Society of Regional Anesthesia (ASRA) guidelines recommend only removing epidural catheters when the INR is less than 1.5. This recommendation is based upon expert opinion and studies suggesting adequate clotting factors are present when the INR is less than 1.5. If the INR is between 1.5 and <3, ASRA recommends cautiously removing epidural catheters only after confirming that the patient is not taking other medications that may affect coagulation studies (i.e, NSAIDS, clopidogrel, low molecular weight heparin, unfractionated heparin). If the INR is >3 ASRA recommends holding or reducing the warfarin dose. No evidence exists to support administration of reversal agents (Vitamin K or FFP) to facilitate removal of epidural catheters. These guidelines may discourage the use of epidural analgesia for postoperative analgesia for total joint replacement.


Methodology This was an observational study of 4,365 patients who had total knee or hip replacement surgery and had their epidural catheters removed with an INR >1.4. A combination of 3,211 patients were prospectively enrolled and 1,154 were retrospectively included in the analysis. Institutions from which the patients were recruited included: the Hospital for Special Surgery (HSS; n = 1617), Rush University Medical Center (RUMC, n=1,594) and Thomas Jefferson University Medical Center (TJU; n = 1,154). Patients undergoing major joint replacement with a requirement for postoperative warfarin administration, and an INR > 1.4 at the time of removal of the epidural catheter were included.


At HSS the epidural catheter was discontinued at noon on POD 2 by the ward nurse if the INR <1.4; this was 36 hours after initiation of warfarin. At TJU epidural catheters are removed in the morning of POD 2. At Rush on POD 3 at 0700 epidural catheters are removed by an acute pain service nurse (removed approximately 50 hours after initiation of warfarin). Neurological checks were performed every two hours for 24 hours after epidural catheter removal in all patients with an INR > 1.4.


Descriptive and inferential statistics were used to analyze the results. The primary outcome was the incidence of epidural hematoma after removal of the epidural catheter with an INR >1.4.


Result In this study, 79% of patients underwent total knee replacement and 21% underwent total hip replacement. The average age was 68 years; the mean duration of epidural analgesia was 2.1 ± 0.6 days. Mean INR at the time of removal was 1.9 ± 0.4, with a range of 1.5 to 7.1. The most frequent INR at which epidurals were removed was 1.6 (n = 1,020 – 23% of patients) followed by an INR of 1.7 (n = 790 – 18% of patients).


In 2% of patients (n = 89) the epidural catheter was removed without complication with an INR >3. At HSS 87% of patients received NSAIDS during warfarin therapy, at RUMC 56% of patients received COX-2 inhibitors, and at TJU all patients received either an NSAID or COX-2 inhibitor in addition to warfarin therapy. No epidural hematomas were observed in any patients (0% incidence, 95% CI: 0% - 0.069%).


Conclusion In this large, multi-center observational study, 4,365 patients taking warfarin had uncomplicated removal of their epidural catheter when the INR was between 1.5 and 7.1. Catheters were cautiously removed with frequent neurological monitoring based on ASRA guidelines.



The most recent ASRA guidelines1 for regional anesthesia in patients receiving antithrombotic or thrombolytic agents were published in 2010. These guidelines were based upon both published data and expert opinion and are designed to improve patient safety. The guidelines form the basis for clinical practice at many facilities when removing epidural catheters in patients taking warfarin who have an INR less than 1.5. Some of the evidence from the ASRA guidelines is lower level evidence, since the incidence of epidural hematoma is extremely rare. This series found no epidural hematomas to occur despite catheters being removed from patients with an INR >1.5. The authors of this study closely followed the ASRA guidelines to ensure the patients were monitored for 24 hours for signs of epidural hematoma.


It is important to point out that the mean INR was 1.9 and that only 2% of patients had an INR >3 (n = 89). I doubt they had a large enough sample size to see an epidural hematoma. Since most of the patients had their epidural catheters safely removed at an INR < 3, I don't think there was enough evidence to support safely removing an epidural catheter with an INR > 3. I would be more conservative than the authors seem to recommend and probably base my practice on current published guidelines. However, in certain situations I might consider removing an epidural catheter between an INR of 1.5 and 1.9. I would ensure hourly neurological monitoring was ordered for at least 24 hours if I removed the epidural catheter for an INR >1.4.


I think the important take home message from this study was that guidelines help inform and guide our practice. Proper clinical judgment and clear communication and departmental/hospital guidelines or procedures are needed to minimize potentially rare, but serious, complications such as epidural hematoma. I believe the increased utilization of acute pain and regional anesthesia services have helped to improve patient safety, and reduce complications in patients taking anticoagulants in the face of neuraxial anesthesia and analgesia.


My guess is the next time ASRA meets the results of this study will be included in the review, and may help inform the new guideline. In the mean time, anesthesia providers can use the current 2010 ASRA guidelines and this study to help support clinical practice.

Dennis Spence, PhD, CRNA

1. Horlocker et al. Executive summary: regional anesthesia in the patient receiving antithrombotic or thrombolytic therapy: Anesthesia Society of Regional Anesthesia and Pain Medicine evidence-based guidelines (third edition). Reg Anesth Pain Med 2010;35:102-105.

The views expressed in this article are those of the author and do not reflect official policy or position of the Department of the Navy, the Department of Defense, the Uniformed Services University of the Health Sciences, or the United States Government.

© Copyright 2011 Anesthesia Abstracts · Volume 5 Number 5, May 31, 2011