ISSN NUMBER: 1938-7172
Issue 1.5

Michael A. Fiedler, PhD, CRNA

Contributing Editor:
Chuck Biddle, PhD, CRNA

Guest Editors:
Mary A. Golinski, PhD, CRNA
Alfred E. Lupien, PhD, CRNA

Assistant Editor
Jessica Floyd, BS

A Publication of Lifelong Learning, LLC © Copyright 2007

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
















Xue FS, Zhang GH, Li XY, Sun HT, Li P, Li CW, Liu KP


Comparison of hemodynamic responses to orotracheal intubation with the GlideScope? videolaryngoscope and the macintosh direct laryngoscope

J Clin Anesth 2007;19:245-250

Xue FS, Zhang GH, Li XY, Sun HT, Li P, Li CW, Liu KP



Purpose            The purpose of this study was to compare the heart rate and systolic, diastolic, and mean blood pressures in patients intubated during general anesthesia with either the GlideScope videolaryngoscope or the Macintosh laryngoscope blade.

Background            Tachycardia and hypertension during endotracheal intubation can be a source of concern in some patients. Stimulation during intubation arises from oropharyngeal and tracheal manipulation. For any given airway manipulation, hemodynamic responses tend to be more severe as the duration of stimulation increases. Different intubation devices may produce more or less tachycardia and hypertension. While the GlideScope videolaryngoscope is used much like a Macintosh laryngoscope blade, the GlideScope may require less upward lifting force to view the glottis since the GlideScope blade is curved 60? and a direct line of sight to the glottis is not required. Exerting less upward force during laryngoscopy may result in less stimulation. On the other hand, the GlideScope requires using an ETT stylet and often requires more time to place the ETT compared to traditional direct laryngoscopy. These factors may increase airway stimulation and the hemodynamic response.

Methodology            This prospective, randomized study included 57 ASA I patients aged 18 to 60 years undergoing elective plastic surgery. Patients taking daily medications that affected blood pressure (BP) or heart rate (HR) and those likely to be difficult to intubate were excluded from the study.

All patients were premedicated with an intramuscular injection of 0.1 mg midazolam and 0.01 mg/kg scopolamine (up to 0.3 mg). Patients were randomized to the GlideScope group or the Macintosh group. General anesthesia was induced with 2 ?g/kg fentanyl, 2 mg/kg propofol, and 0.1 mg/kg vecuronium. All intubations were performed by the same individual. GlideScope intubations were performed using an ETT with a lubricated stylet bending the ETT 60? to the shape of the GlideScope blade according to manufacturer’s recommendations. Macintosh intubations were performed without a stylet under direct visualization. Males (n=20) were intubated with a 7.5 mm endotracheal tube (ETT) and females (n=37) with a 7 mm ETT. Anesthesia was maintained with 1% isoflurane, 60% nitrous oxide, and normocapnia.

BP and HR were recorded after induction, at intubation, and at one minute intervals for the next 5 minutes. The duration of intubation was also recorded. Patients who required more than one attempt at intubation were excluded from analysis.

Result            Two GlideScope patients required more than one laryngoscopy and were excluded from analysis.

Average intubation time was 37.4?9.9 sec in the GlideScope group and 28.4?11.7 sec in the Macintosh group (P<0.01). Intubation time was 24% longer in the GlideScope group.

Intubation resulted in statistically significant increases in BP and HR in both groups. The average maximum HR was 24 bpm higher than the pre-intubation HR in GlideScope patients and 15 bpm higher in Macintosh patients. GlideScope HR increases lasted an average of 4 minutes while HR increases in Macintosh patient lasted an average 1 minute.

The average maximum systolic BP was +5 torr and -1.5 torr compared to pre-intubation systolic BP in GlideScope vs. Macintosh patients respectively. There were no statistically significant differences in BP or HR between the GlideScope and Macintosh groups at any time.

Conclusion            Changes in BP and HR were similar following GlideScope and Macintosh laryngoscopy. This was true despite the use of a stylet and a longer intubation time in the GlideScope group.



This study is a good first step in answering the question, “is the GlideScope more or less stimulating than traditional direct laryngoscopy?” In this study only patients with normal airways were included so the investigators were comparing the Macintosh blade in a situation where it would normally be used with the GlideScope in a situation where it normally would not be used. Also, all the patients were healthy and would not be expected to have severe hemodynamic responses to laryngoscopy and intubation. Both these factors make the comparisons of limited value, but, again, it is a place to start.

The changes in heart rate and blood pressure seen in this study were minimal, as one would expect in healthy patients. The largest hemodynamic change was a 24 bpm increase in HR in the GlideScope group. This increase in HR might well be clinically significant in some circumstances. It would be difficult to argue that any of the other changes in HR or BP were clinically significant. Even more importantly, the differences in HR and BP between groups weren’t even statistically significant. In this respect, the GlideScope may be seen to be the winner since, clinically, hemodynamics were similar to the Macintosh group despite a longer time to intubation and the use of a stylet. The one caveat here is that the heart rate increase in the GlideScope group lasted an average of 4 minutes while the HR increase in the Macintosh group lasted only 1 minute. It would be interesting to know how long heart rate increases would persist following a GlideScope intubation in a poorly controlled hypertensive patient with a difficult airway.

Lastly, while these investigators seemed to focus on the stimulation of laryngoscopy, the most stimulating part of endotracheal intubation is inflating the cuff on the ETT. Since one must inflate the cuff regardless of the method used to place the ETT I have to wonder how important the hemodynamic changes associated with different intubation devices really are.

Michael Fiedler, PhD, CRNA




© Copyright 2007 Anesthesia Abstracts · Volume 1 Number 5, August 31, 2007

Baraka AS, Taha SK, Siddik-Sayyid SM, Kanazi GE, El-Khatib MF, Dagher CM, Chehad J-M A, Abdullah FW, Hajj RE

Supplementation of pre-oxygenation in morbidly obese patients using nasopharyngeal oxygen insufflation

Anaesthesia 2007;62:769-773

Baraka AS, Taha SK, Siddik-Sayyid SM, Kanazi GE, El-Khatib MF, Dagher CM, Chehad J-M A, Abdullah FW, Hajj RE


Purpose            The purpose of this study was to evaluate the effectiveness of supplemental nasopharyngeal oxygen insufflation for delaying desaturation in obese patients during induction of general anesthesia.

Background            Following preoxygenation, desaturation occurs more quickly during apnea in obese than non-obese individuals. Preoxygenation in the sitting or head elevated position has been recommended for obese patients. Insufflating oxygen into the pharynx during apnea has been shown to slow desaturation. Morbidly obese patients desaturate quickly during apnea due to a reduced functional residual capacity (FRC) and increased oxygen consumption. The supine position and induction of general anesthesia also decreases FRC. FRC decreases over twice as much in obese compared to non-obese patients with the induction of general anesthesia. A sufficient decrease in FRC results in tidal volume breathing that encroaches upon the FRC and results in small airway closure during exhalation.

During apnea, oxygen is removed from the lungs by the circulation. The rate of oxygen uptake is variable but one study reported it to be approximately 250 mL/min. Carbon dioxide is quite soluble in blood. In the same study, CO2 entered the lungs at a rate of only about 10 mL/min. This resulted in a net loss of 240 mL/min of gas from the lungs and an intrapulmonary pressure that was lower than atmospheric pressure. Given a patent airway, 240 mL/min of gas would thus be entrained into the lungs. Pharyngeal insufflation of oxygen during induction of anesthesia may benefit morbidly obese patients.

Methodology            This prospective, randomized, double-blind study included 34 ASA class I and II patients with a Body Mass Index (BMI) >35 kg/m2 scheduled for elective gastric bypass or gastric banding. Eligibility criteria included a Mallampati class I or II airway, a thyromental distance ≥ 6 cm, and the absence of cardiopulmonary pathology. Patients were examined to verify patency of their nasal passages.

All patients were preoxygenated with 100% oxygen via an anesthesia mask for three minutes with the head of the bed elevated 25?. Anesthesia was induced with midazolam 2 mg, fentanyl 2 ?g/kg, propofol 2 mg/kg, and succinylcholine 2 mg/kg. Cricoid pressure was applied. When breathing stopped, a 10 French catheter was inserted through the nose and positioned in the nasopharynx of all patients. In the study group, oxygen 5 L/m was insufflated through the nasopharyngeal catheter. In the control group, no supplemental oxygen was administered via the nasopharyngeal catheter. Next, an experienced laryngoscopist performed a direct laryngoscopy in all patients without intubating the trachea. The laryngoscopy and cricoid pressure were continued for the duration of the study period. Patients remained apneic until their oxygen saturation decreased to 95% or four minutes elapsed. When either the saturation fell to 95% or four minutes was reached all patients were endotracheally intubated and ventilated with 100% oxygen. The scheduled surgery then proceeded.

Result            BMI was comparable in the study and control groups. All patients reached 100% oxygen saturation during the preoxygenation period. End tidal oxygen concentrations averaged 89% in both the study and control groups at the end of preoxygenation.

During apnea, control group oxygen saturation decreased to 95% in an average of 145 (sd 27) seconds (2 minutes, 25 seconds). All control patients desaturated to 95% within the four minute study period. Control patients exhibited a correlation between the time required for their saturation to fall to 95% and BMI (r2=0.66, P<0.05). The higher their BMI, the faster they desaturated.

In the study group, oxygen saturation remained 100% during the four minute data collection period in 16 of 17 study patients. The 17th study patient had a BMI of 65 kg/m2 and desaturated to 95% in 153 seconds.

Conclusion            Delivering supplemental nasopharyngeal oxygen to morbidly obese patients during the apnea associated with induction of general anesthesia and intubation significantly delays the onset of arterial desaturation.



I was introduced to the concept of apneic oxygenation years ago by a colleague but I’ve not seen it put to good use in anesthesia until now. The 1959 study1 I am familiar with is one of those truly enlightening studies that doubtless could not be performed today. In it, eight surgical patients were paralyzed, intubated, and ventilated with 100% oxygen for 30 minutes. In some cases the study period was prior to the surgical procedure and in other cases it was during surgery. After 30 minutes of oxygenation the patients, still paralyzed, were allowed to remain apneic with their endotracheal tube connected to the anesthesia circuit containing 100% oxygen. The periods of apnea ranged from as little as 18 minutes to a high of 55 minutes. Arterial blood gases were drawn every five minutes. The lowest oxygen saturation was 98% in two patients apneic for 45 and 53 minutes respectively. The highest PaCO2 was 250 torr in a patient apneic for 45 minutes. In most patients the highest PaCO2 was 130-160 torr. The elevation in PaCO2 resulted in the expected respiratory acidosis with a pH as low as 6.72 after 53 minutes of apnea. Ventilation was restarted in two patients with the onset of cardiac arrhythmias. Ventricular arrhythmias did not occur in the other six patients. All patients survived. This, and other studies like it from the same time period, demonstrate an important point. Healthy patients can be adequately oxygenated for far longer than one would think if they have a patent airway and a high concentration of oxygen available for mass movement into the lungs as oxygen is taken up by the circulation.

Perhaps we should consider having oxygen catheters and an oxygen source separate from our anesthesia circuit available at all anesthetizing locations. In patients at risk for desaturation, such as the morbidly obese patients included in this study, a pharyngeal oxygen catheter could be used routinely to increase the margin of safety. When an unexpected difficult airway is encountered the first thing we do is insert the oxygen catheter into the pharynx and crank up the flow. When a difficult airway is anticipated, in addition to calling for the difficult airway cart, we have the pharyngeal oxygen catheter ready to insert. If the results of this study are a true indication, doing so would buy us a lot more time to get an endotracheal tube in place.


Michael Fiedler, PhD, CRNA


1.            Frumin MJ, Epstein RM, Cohen G. Apneic oxygenation in man. Anesthesiology 1959;20:789-97.



© Copyright 2007 Anesthesia Abstracts · Volume 1 Number 5, August 31, 2007


Ezri T, Weisenberg M, Sessler DI, Berkenstadt H, Elias S, Szmuk P, Serour F, Ebron S


Correct depth of insertion of right internal jugular central venous catheters based on external landmarks: avoiding the right atrium

J Cardiothorac Vasc Anesth 2007;21:497-501

Ezri T, Weisenberg M, Sessler DI, Berkenstadt H, Elias S, Szmuk P, Serour F, Ebron S


Purpose            The purpose of this study was to compare the number of Central Venous Catheters (CVCs) requiring repositioning following insertion to a standard depth vs. a depth calculated from measuring the distance between two surface landmarks.

Background            FDA guidelines recommend against positioning the tip of a central venous catheter in the right atrium. Right atrial placement can result in perforation and cardiac tamponade. Incorrect positioning of CVCs requiring reinsertion or repositioning is common. By some measures up to 30% of CVCs are positioned with the catheter tip in the right atrium. Ideal CVC insertion depth from a right internal jugular (IJ) approach has been recommended to be 16.5 cm and less than 15 cm. Double lumen CVCs currently available are up to 20 cm long. Radiographically, the junction of the superior vena cava and the right atrium is near the angle between the trachea and right mainstem bronchus. When seen on x-ray, the proper CVC insertion depth results in the tip of the catheter lying near the carina. Some have recommended using external landmarks to achieve this placement. Using surface landmarks to determine the length of CVC to insert during right IJ central line placement may reduce the need for repositioning of CVCs and reduce complications.

Methodology            This prospective, randomized, double-blind study included 100 consecutive patients scheduled for laparotomy who required a central line. All patients had a 7.5 French, double-lumen CVC inserted via the right IJ vein. In the 15 cm group the CVC was inserted to a fixed depth of 15 cm. In the Measured group the CVC was inserted to a depth determined by measuring the distance between two external landmarks. The distance between two points was measured to determine how far to insert the CVC in this second group. The first point was between the carotid pulsation and the medial border of the sternocleidomastoid muscle on a transverse line crossing the superior border of the thyroid cartilage. The first point was also the CVC insertion site for both groups. The second point was on the right lateral boarder of the manubrium at the junction between the upper and middle thirds of the manubrium. The actual location of the tip of the CVC was determined with an anteroposterior chest x-ray (CXR) taken in a standardized manner. The radiologist interpreting the CXR did not know which group the patient was in. The CVC was defined as needing repositioning if the catheter tip was > 5 cm above the carina or > 1 cm below the carina.

Result            In the Measured group, CVCs were inserted to depths between 9 cm and 12.5 cm. No CVCs in either group were positioned with the catheter tip > 5 cm above the carina. In the 15 cm group, 78% of catheter tips were > 1 cm below the carina compared to 2% (1 of 50 patients) in the Measured group (P<0.001). The tip of the catheter was in the right atrium in 20% (10 of 50 patients) of the 15 cm group and none of the Measured group (P<0.05). All catheter tips in the 15 cm group were below the carina; they averaged 1.9 cm below the carina. The tip of the Measured catheters averaged 2.9 cm above the carina. One CVC in the measured group and 39 CVCs in the 15 cm group required repositioning (P<0.001).

Conclusion            Inserting a right IJ CVC to a depth calculated by measuring the distance between surface landmarks takes into consideration interindividual variability and reduces the need to reposition CVCs after obtaining a CXR.



This is one of those “puzzle solving” studies that makes all kinds of sense and has the potential to really improve one little area of our practice. Even if there are no complications involved, having to go back and reposition almost 80% of the central lines inserted is quite time consuming. Even if I’m not the one who has to do it, someone has to do it and that affects the group’s overall workload. In this case, I’d say the investigators guessed a little short in the measured group. We might want to calculate a slightly longer length when we try this technique. And, of course, their measurement is from a very specific insertion point. A slightly higher or lower central line insertion point would require an equal adjustment in the length of the catheter inserted. I suspect that most anesthetists will have to “fine tune” the external landmarks they use to come up with an insertion depth that positions the tip of the CVC in an acceptable location based upon their usual insertion site. That shouldn’t be too hard to do. I expect this little procedure may be genuinely helpful.


Michael Fiedler, PhD, CRNA

© Copyright 2007 Anesthesia Abstracts · Volume 1 Number 5, August 31, 2007

Orme RML?E, McSwiney MM, Chamberlin-Webber RFO



Fatal cardiac tamponade as a result of a peripherally inserted central venous catheter: a case report and review of the literature

Br J Anaesth 2007;99:384-388

Orme RML’E, McSwiney MM, Chamberlin-Webber RFO




Purpose            The purpose of this article was to report and discuss a fatal cardiac tamponade involving a PICC central venous catheter.

Background            Central lines may be placed from distal sites, such as the arm. Peripherally Inserted Central venous Catheters (CVC) are called PICC lines. PICC lines are often used for long term central venous access and may be used by anesthesia during, for example, head and neck surgery. Cardiac tamponade is a known complication associated with central venous catheters and is often fatal. It has been reported when PICC lines are used in pediatrics but has been rarely reported in adults. Risk factors include the position of the tip of the catheter and the material the catheter is made out of (Editor’s Note: this may actually address the stiffness of the catheter). Catheter tip position may be hard to control with a PICC line because changes in arm position have been shown to result in tip movement of several centimeters.

The optimal position of the tip of a CVC is not agreed upon. The FDA states that the tip of a CVC should not lie within the heart. Other experts believe that the tip of a CVC may safely lie within the right atrium, with some limitations. Case reports, however, seem to identify intra-atrial placement as a risk factor for cardiac tamponade. When a central line is in place and a cardiac tamponade is suspected, fluid flow through the central line should be stopped and the catheter aspirated.

Methodology            A 20 year old with asthma underwent an elective maxillary osteotomy with general anesthesia. In addition to routine monitors, an arterial line and central line were placed. The central line was a 1.7 mm external diameter polyurethane PICC line, inserted at the right antecubital fossa. The PICC line was inserted and advanced without difficulty. After insertion, the guidewire was used to estimate the position of the tip of the PICC line. Venous blood was aspirated and a characteristic central venous waveform appeared when the line was transduced. The PICC line was transduced continuously throughout the case. It was not used intraoperatively for fluid administration.

During the 8 hour procedure the patient received 3.5 L of crystalloid and 1.5 L of colloid. Postoperatively, the patient was sedated and ventilated overnight. A chest x-ray was taken after arrival in the ICU. The tip of the PICC line was visible in the right atrium. The PICC line was not withdrawn. At about 9 am the next morning, the patient was extubated, received a unit of packed red blood cells, and had a potassium of 3.5 mmol/L. Potassium was infused through the PICC line over 2 hours. Thirty minutes after the potassium infusion was finished ST elevation was noted on the ECG monitor. The patient complained of shoulder pain. A 12-lead ECG showed ST elevation in leads II, III, V1, and V2. Thirty minutes later, blood pressure fell to 80/50. A bolus of 200 mL hydroxyethyl starch was administered via the PICC line. BP suddenly fell to 40 systolic and the patient became unconscious. Resuscitation was begun and periods of circulatory arrest ensued. Total resuscitation time was 35 minutes. During the resuscitation, pericardiocentesis was performed twice and a total of 220 mL of fluid withdrawn. A perfusing rhythm returned after pericardiocentesis. Immediately after resuscitation venous blood could be withdrawn from the PICC line. A post-arrest x-ray showed the tip of the PICC catheter in the superior vena cava.

Result            Post resuscitation, the patient’s pupils were fixed and dilated. CT scan showed severe hypoxic brain injury. An autopsy was performed. No signs of cardiac perforation by the PICC line were found. The PICC catheter showed no signs of damage or defect. Analysis of the fluid aspirated during pericardiocentesis produced a potassium concentration of 36.4 mmol/L and concentrations of sodium and chloride consistent with the normal saline the potassium was diluted in prior to administration through the PICC line. It appeared that the PICC line had migrated into the pericardial space. Potassium was infused into the pericardial space where it diffused into the myocardium causing cardiac arrest. The volume of the potassium solution and/or the bolus of hydroxyethyl starch caused the cardiac tamponade and resulted in the initial deterioration in vital signs.

Conclusion            Fluid was infused through the PICC line into the pericardial space resulting in cardiac arrest and, ultimately, brain death. Central venous catheter tip position should be verified by x-ray before fluid is infused through a CVC. Cardiac tamponade should be considered in any patient with deteriorating vital signs who has a CVC in place. In such a case, the CVC should be aspirated.



I suspect that fear of litigation might have prevented this case report from being submitted had the case occurred in the USA. That is unfortunate, because there is much to learn from reports such as this. The best evidence I have read indicates that the tip of a central venous catheter should not be allowed to lie in the heart. The carina is an agreed upon landmark useful for judging the position of the tip of a central venous line on an x-ray. It lies near the junction of the superior vena cava and the right atrium. Another abstract in this issue (Correct depth of insertion of right internal jugular central venous catheters based on external landmarks: avoiding the right atrium, J Cardiothorac Vasc Anesth 2007;21:497-501) provides evidence that the tip of the central line should be visible no more than 1 cm below the carina. Given how difficult it is to properly position the tip of a PICC line by measuring the length of catheter inserted, it seems reasonable to obtain a chest film and reposition the catheter early on. In hindsight, it would have been nice to have had that chest x-ray before starting an 8 hour case. If the catheter tip location was the cause of the complication in this case (and we don’t know for certain that it was) there was another chance to withdraw the catheter tip back into the vena cava when an x-ray was taken in the ICU.

We all see central lines used frequently without problems. But when central line complications occur they can be quite severe. This case report should serve to remind us of the risk associated with central lines, the importance of verifying proper placement, and the need for maintaining a high index of suspicion when a central line is in place. Aspirate frequently. When in doubt, use another line.


Michael Fiedler, PhD, CRNA


Additional articles validating the use of the carina as a landmark for positioning the tip of central venous catheters.


Albrecht K, Nave H, Breitmeier D et al. Applied anatomy of the superior vena cava-the carina as a landmark to guide central venous catheter placement. Br J Anaesth 2004;92:75-7.

Stonelake PA, Bodenham AR. The carina as a radiological landmark for central venous catheter tip position. Br J Anaesth 2006;96:335-40.



© Copyright 2007 Anesthesia Abstracts · Volume 1 Number 5, August 31, 2007

Equipment & Technology

Lui N, Chazot T, Gentry A, Landais A, Restoux A, McGee K, Lalo? P, Trillat B, Barvais L, Fischler M



Titration of propofol for anesthetic induction and maintenance guided by the bispectral index: Closed-loop versus manual control

Anesthesiology 2006;104:686-695

Lui N, Chazot T, Gentry A, Landais A, Restoux A, McGee K, Laloë P, Trillat B, Barvais L, Fischler M




Purpose            To compare the performance of closed-loop and manual target-controlled infusion (TCI) systems for administration of propofol during the induction, maintenance, and emergence phases of general anesthesia.

Background            Aircraft pilots use automated systems to reduce workload and minimize the potential for human error. Similarly, the quality of anesthesia potentially can be improved by using target-controlled infusions of propofol titrated to the bispectral index (BIS) by a microprocessor using a proprietary proportional-differential algorithm.

Methodology            One hundred sixty-four adult patients, ASA I-III,  presenting for elective surgical procedures longer than 30 minutes were randomized to receive propofol and remifentanil by TCI adjusted either by microprocessor or manually to maintain a BIS of 50. Propofol and remifentanil infusions were discontinued when the surgical procedure was complete. Outcome variables included duration of induction, time from discontinuation of infusion to extubation, percentage of time BIS remained between 40 and 60, overshoot (BIS<40) and undershoot (BIS>70), performance error (the difference between actual and intended BIS), wobble (within-subject variation in performance error), and a global score computed as the ratio between median performance error plus wobble, and percentage of time with BIS between 40 and 60.

Result            There were no significant differences between the closed-loop and manually-controlled groups ages, sex, height, weight, or severity of surgical procedures. Despite random assignment, there were more ASA III patients in the closed-loop group (16% versus 4%).

There were no inter-group differences in premedication, remifentanil doses, overshoot, or uses of ephedrine or antihypertensive medications during the induction phase of anesthesia. The mean induction time was longer in the closed-loop group (320 versus 271 s); however the closed-loop group also required a lower propofol induction dose (1.4 versus 1.8 mg/kg), achieved a lower propofol effect-site target dose (3.2 versus 3.8 ?g/mL), and experienced less BIS overshoot (12 versus 29 s).

During the maintenance phase, there were no differences between groups in duration of anesthesia, weight-adjusted doses of propofol or remifentanil, weight-adjusted volume of Ringer’s lactate administered, or percentage of patients requiring either ephedrine or antihypertensive therapy. Statistically significant differences were noted in mean time to tracheal extubation (7 min in the closed-loop group compared with 11 min in the comparison group) and patients with blood loss greater than 500 ml (22% in the closed-loop group compared with 7% in the comparison group). There were also statistically significant inter-group differences in propofol target adjustments:  the closed-loop group had more frequent (33 versus 11 per hour) and smaller (.30 versus .69 ?g/mL) adjustments compared with the manual TCI group. Only 4 of 5,273 target adjustments in the closed-loop group were made manually compared with 1,543 manual adjustments in the comparison group. The target BIS level between 40 and 60 was maintained 89% (?9%) of the maintenance time in closed-loop patients compared with 70% (?21%) in the comparison group. There was no difference in undershoot between the two groups; however, overshoot was observed more frequently in the manual TCI group (26?22%) compared with the closed-loop group (8?8%). There were no incidents of awareness reported by patients in either group. Performance error, wobble and global scores were all significantly higher in the manual TCI group.

Conclusion            Closed-loop TCI of propofol titrated to BIS resulted in more frequent and finer adjustments of propofol delivery with less variation in drug level. Manually-controlled TCI was more efficient during induction of anesthesia. During the maintenance and emergence phases of general anesthesia, the closed-loop system reduced overall consumption of propofol, increased the percentage of time BIS was maintained between 40 and 60, reduced the amount of time with BIS less than 45, and decreased emergence time in adult surgical patients. Hemodynamic stability was equivalent between groups even though anesthesia providers controlling the comparison group TCI manually had access to each patient’s history and concurrent hemodynamic data whereas dose adjustments in the closed-loop group were made exclusively based on BIS (without consideration of cardiovascular parameters). In addition, there was no difference in hemodynamic stability between groups despite a higher percentage of ASA III patients and greater blood loss occurring in the closed-loop group. Closed-loop TCI of propofol, titrated to BIS, can be safe, reliable, and feasible; and potentially decrease anesthesia provider workload.



In his book, The Logic of Failure1, Dietrich Dorner describes a series of experiments exploring how we humans fail at various tasks. One of the experiments required participants to manually regulate the thermostat of a dairy cooler to maintain a narrow temperature range and prevent the cooler’s products from spoiling. Most participants were not very successful. In some circumstances, thermostat changes were too gradual, and in others, too exaggerated. Dorner’s experiment seems very analogous to our task of titrating anesthetic and vasoactive agents, except that broad swings in hemodynamic or anesthetic levels can be costly, both physiologically and economically. One potential solution for us as anesthesia providers would be to use a microprocessor “thermostat” to automate the task of anesthetic titration.

Target-controlled infusion (TCI) is a term used to describe the use of a computer to control the rate of administration of an anesthetic agent to achieve a pre-determined endpoint. Unlike a traditional infusion of agent at a consistent rate, the rate of a TCI varies over time based on the specific agent’s pharmacokinetics and patient parameters. The ‘simpler’ version of TCI uses an algorithm to estimate infusion rates to achieve the desired blood, plasma, or effect site concentration of the anesthetic. More complex TCI models adjust drug concentrations based on closed-loop feedback of a physiological variable, such as either EEG or BIS. Liu et al proposed that reducing the amount of time a clinician spends adjusting infusions would conserve time that the provider could use to concentrate on other anesthesia tasks, and they offer some indirect evidence to support their claim: this study’s experimental group maintained similar hemodynamic profiles compared to the comparison group, despite having higher acuity patients and greater blood losses. But, the current investigation uses simple TCI in the comparison group; consequently, we’re forced to reach conclusions about differences between closed-loop TCI and traditional (non-TCI) titration of anesthetic agents through inference from older studies comparing simple TCI with manual titration. In any case, it seems reasonable to conclude that there is emerging evidence that computer-controlled infusions may assist us by reducing our overall cognitive and physical workload. As one minor criticism of the investigation, it would have helpful if the authors reported the anesthesia providers’ workload (perhaps using a survey instrument as simple as the NASA Task Load Index) or frequencies of errors or near misses, if any, since their fundamental premise was that closed-loop TCI improves anesthesia quality by reducing provider workload.

Even with the many positive findings from this investigation, controlled-loop TCI is not ready for widespread implementation. Induction of anesthesia was slower than manual or simple TCI and both the simple and feedback approaches to TCI, when out of the target range, were more likely to overshoot than undershoot anesthetic depth. It is not surprising that overshoot would be the more common variance because rising BIS levels are easily and rapidly offset by increasing the rate of agent administration while the effects of excessive dosing can be terminated only by agent metabolism. Although one would be inclined to believe that a little too much anesthesia is acceptable as long as the patient remains hemodynamically stable, the investigators reference a 2003 abstract by Lenmarken, Lindholm, Greenwald, and Sandler who reported that low BIS levels were associated with increased patient mortality.2 What puzzles me the most, is that the authors did not also cite the more recent and extensive observational study of 1064 cases by Monk, Saini, Weldon, and Sigl who reported that “cumulative deep hypnotic time” of BIS < 45 was associated with a 24% increase in one-year patient mortality.3 For me, the study of Monk et al is particularly noteworthy because it was the impetus for an editorial and a series of letters to the editor of Anesthesia and Analgesia expressing concern regarding the investigators’ interpretation of their findings. The most succinct criticism of Monk et al’s conclusions was advanced by Arnold J. Berry who cautioned readers that the observed statistical association between BIS level and mortality does not provide evidence of causation.4  Looking beyond the controversy between low BIS and patient mortality, it will be interesting to observe how the role of closed-loop TCI evolves over the coming years as proprietary decision algorithms become commercialized and evidence accumulates regarding TCI’s safety and effectiveness.



Alfred E. Lupien, PhD, CRNA


1. Dorner, D. (1996). The logic of failure: Recognizing and avoiding error in complex situations (Translated by Rita and Robert Kimber). Cambridge, MA :Perseus 

2. Lenmarken C, Lindholm MJ,Greenwald SD, Sandler, R. Confirmation that low intraoperative BIS levels predict increased risk of post-operative mortality (abstract). Anesthesiology 2003; 99:A303

3. Monk TG, Saini V, Weldon BC, Sigl JC. Anesthetic management and one-year mortality after noncardiac surgery. Anesth Analg  2005; 100:4-10

4. Berry AJ. Observational studies identify associations, not causality (Letter). Anesth Analg 2005; 101:1238


© Copyright 2007 Anesthesia Abstracts · Volume 1 Number 5, August 31, 2007


Bendavid E, Kaganova Y, Needleman J, Gruenberg L, Weissman JS


Complication rates on weekends and weekdays in US hospitals

Am J Med. 2007;120:422-428

Bendavid E, Kaganova Y, Needleman J, Gruenberg L, Weissman JS


Purpose            To compare weekday versus weekend hospital complication rates in surgical and obstetrical patients.

Background            Several factors converge to increase the potential for patient complications during weekends: there are fewer experienced physicians and nursing personnel, proportionally more patients are admitted through the emergency department, there are fewer patient discharges, and some hospital services may not be available. Results from research attempting to determine whether there are, in fact, different mortality rates have been mixed. Interpretation of mortality data is difficult because the cause of a fatal outcome can not be consistently attributed to a specific day of the week; however, surgical complications (such as retained foreign bodies, laceration, and hemorrhage) can be matched with the specific date and time of the procedure.

Methodology            Administrative records containing data from all inpatients hospitalized between 1999 and 2001 in New York and Massachusetts, plus North Carolina data from 2000-2001 were obtained through the Healthcare Cost and Utilization Project for analysis. Available data included demographics, date and route of admission, dates and ICD-9-CM codes for all diagnoses and procedures, birth information, and discharge status. From the approximately 20 potential Patient Safety Indicators developed through the Agency for Healthcare Research and Quality, eight complications were identified where there was potential to isolate the precise date of complication. The indicators included anesthesia complications; retained foreign body, postoperative hemorrhage, and accidental laceration in surgical patients; obstetrical trauma to the parturient during Cesarean section and vaginal delivery (with and without instrumentation), and infant birth trauma. “Weekend” was defined as the period between midnight on Friday evening and midnight on Sunday evening. Data were analyzed using logistic regression. Patient demographics and comorbidities, as well as admission type and route were used to adjust the odds ratios. An additional analysis was conducted with patients requiring selected vascular (including carotid endarterectomy and femoropopliteal bypass) and cardiac (such as coronary artery bypass grafting and pacemaker insertion) procedures to determine whether patients requiring highly specialized or complex surgery would be more vulnerable to differences between weekends and weekdays.

Result            Data were collected for 4,967,114 admissions of patients identified as being at risk for at least one of the study complications. Patients admitted on weekends represented 14.8% of the sample and tended to be younger and non-white.

Overall, 114,090 complications were detected in 2.3% of the admitted patients. Of the complications, 66.0% were obstetrical, 28.3% surgical (including anesthesia), and 5.7% neonatal. The overall rate of anesthesia complications was 61 complications (per 100,000 patients), representing 1.6% of all complications, with an adjusted rate of 54 on weekends and 63 on weekdays. The adjusted weekend:weekday odds ratio (OR) was 0.86. Anesthesia-related complications was the only category where complications were less frequent on weekends than weekdays.

The other statistically significant complication rate differences were for postoperative hemorrhage (OR = 1.07), vaginal delivery without instrumentation (OR = 1.03), Cesarean section (1.36), and birth trauma (OR = 1.06). There was a statistically significant increase in complications for vascular surgery patients (OR = 1.46), but not for cardiac surgery patients. No analyses of anesthesia complications were reported for the subsets of vascular and cardiac surgical patients.

Conclusion            There was a small increase in the rates of some complications for individuals receiving acute inpatient care on weekends. The weekend effect was most pronounced for patients requiring Cesarean section or vascular surgery, and is potentially the result of hospital staffing patterns and resource utilization.

The weekend decrease in anesthesia complications is “interesting” and potentially attributable to the leadership of anesthesiologists in reducing risk through the identification of factors such as communication failure and production pressure. Potentially, strategies to minimize these factors are more effective during weekends when surgical volume is reduced.


Statistical analysis of large data sets using logistic regression often leads to findings that are both interesting and easy to interpret, through the reporting of odds ratios. Oftentimes, the data used for these analyses were collected for administrative purposes (such as record keeping and billing) rather than the study of clinical outcomes; consequently, the research team must be very careful to control for any bias that is inherent to the data set, and exercise caution when operationally defining variables to avoid misinterpretation of results. Accepting these limitations, Bendavid et al report findings with implications for us as anesthesia providers. Because we often find ourselves managing the clinical consequences of surgical and obstetrical complications, or we are included in litigation following a complication, recognition of the potential for higher weekend rates of some surgical and obstetrical complications can serve as an additional impetus for our clinical vigilance and thorough documentation.

I agree with the investigators that a lower rate of anesthesia complications on weekends is intriguing. Perhaps the involvement of fewer anesthesia providers with an individual patient’s care on weekends reduces morbidity by eliminating many opportunities for communication lapses from the period of pre-anesthetic assessment through PACU transfer; however, I suspect that the explanation for any rate difference is more complex than the authors’ conjecture about anesthesiologist-developed processes for mitigating production pressure and communication lapses. First, without an established ‘baseline’ for anesthesia complication rates, it is impossible to determine whether complication rates were 14% lower on weekends, or higher on weekdays. Even if we accept the authors’ explanation, any difference in complication rates may indicate that the anesthesia community hasn’t been as effective as we can be in managing production pressure etc. Secondly, unlike the clinical practices of surgeons and obstetricians which are fundamentally unchanged between weekdays and weekends, models for anesthesia delivery often differ between weekdays and weekends. During weekend and evening hours: supervisory ratios in the care team model may vary, patient care may shift from a team model to personally-administered, personally-administered care may shift from MD to CRNA (or vice versa), or the involvement of physician or nurse residents may change. Any combination of these factors could contribute to a difference in complication rates.

I’m reminded of the axiom that we CRNAs magically become more capable after 5:00 p.m. and on weekends; but as a scientist, all I can genuinely conclude from the findings of Bendavid et al is that something appears to be different about how anesthesia care is administered on weekends. It would be interesting to tease out some of these answers by exploring anesthesia staffing patterns.


Alfred E. Lupien, PhD, CRNA

A weekend:weekday odds ratio (OR) of 1.07 for postoperative hemorrhage indicates that hemorrhage was 7% more likely on a weekend, i.e. there were 107 incidences of weekend postoperative hemorrhage for ever 100 weekday incidences. Similarly, an OR of .86 for anesthesia complications indicates that there were 86 weekend anesthesia complications for every 100 weekday complications, or 14 per 100 (14%) less.

© Copyright 2007 Anesthesia Abstracts · Volume 1 Number 5, August 31, 2007

Obstetric Anesthesia

Dahlgren G, Granath F, Wessel H, Irestedt L


Prediction of hypotension during spinal anesthesia for cesarean section and its relation to the effect of crystalloid or colloid preload

Int J Obstet Anesth 2007;16:128-134

Dahlgren G, Granath F, Wessel H, Irestedt L



Purpose            The purpose of this study was to compare the incidence of hypotension in term pregnant women undergoing subarachnoid block for cesarean section in relation to the results of their Supine Stress Test (positive or negative) and their receipt of either a crystalloid or colloid IV preload.

Background            Aortocaval compression poses a risk to fetal oxygenation. It produces symptoms detectable by the parturient in about 8% of term pregnant women. A larger portion of women may experience aortocaval compression yet perceive no symptoms. Subarachnoid block anesthesia may worsen maternal hypotension when combined with the effects of aortocaval compression. The ability to detect women at risk for aortocaval compression might also predict those at greatest risk for hypotension following subarachnoid block. The preoperative supine stress test (SST) was shown to predict severe maternal hypotension in one study. Intravenous fluid preloading with a colloid solution may more effectively prevent hypotension following subarachnoid block for cesarean section than equal volumes of crystalloid solutions.

Methodology            This randomized, double-blind study included 53 healthy, singleton, term pregnant women scheduled for elective cesarean section with subarachnoid block anesthesia. Each woman underwent a SST preoperatively. The SST involved recording the fetal heart rate (FHR), maternal blood pressure (BP) and maternal heart rate (HR), maternal sensations, and maternal behaviors while lying left lateral and while lying supine. The SST was positive if lying supine resulted in a sustained increase in HR of >10 bpm, a sustained decrease in BP of >15 torr, reports of nausea or dizziness, or certain observable maternal behaviors.

Women were randomized to receive either 1000 mL Ringer’s solution (crystalloid group) or 1000 mL 3% Dextran 60 (colloid group) over 20 minutes immediately prior to induction of their subarachnoid block. Hyperbaric bupivacaine 12.5 mg with 10 ?g fentanyl was administered with patients in the sitting position. Immediately after the block was injected parturients were positioned supine with left uterine displacement. All women received supplemental oxygen by nasal cannula.

Hypotension was defined as a systolic BP below 100 torr or maternal symptoms of nausea, vomiting, dizziness, or chest symptoms accompanies by a systolic BP at least 20% below baseline. Clinically significant hypotension was defined as hypotension accompanied by nausea, vomiting, dizziness, or chest symptoms. Ephedrine was administered by protocol for hypotension but IV fluid flow was not increased in response to hypotension.

Thus, women were first divided into groups by those with a positive SST and those with a negative SST. They were then further divided into subgroups of those who received crystalloid IV fluid preloading and those who received colloid IV preloading.

Result            Of the 53 women included in the study, 36% (19) had a positive SST and 64% had a negative SST. Of the 19 with a positive SST, 16 experienced an elevated HR and 3 experienced hypotension. Many also met sensory or behavioral criteria for a positive SST as well. All women  who underwent the SST had normal FHR tracings throughout the SST.

Clinically significant hypotension occurred in 68% of the women in the positive SST group (both crystalloid and colloid subgroups) compared to 29% of women in the negative SST group. Severe hypotension occurred in 32% of the women in the positive SST group (both crystalloid and colloid subgroups) compared to 6% of women in the negative SST group.

Within the positive SST group, the total dose of ephedrine administered was greater (20 mg vs. 7 mg) in women who received a crystalloid preload than in those who received a colloid preload (P=0.01). Clinically significant hypotension was more common (P=0.003) and the total dose of ephedrine administered was greater (P=0.002) in the positive SST / crystalloid women than in all other subgroups pooled together (positive SST / colloid, negative SST / crystalloid, negative SST / colloid).

All 5 minute APGAR scores were ≥9. The SST had a positive predictive value of 90% for hypotension in healthy term pregnant women.

Conclusion            Healthy, singleton, term pregnant woman with a positive SST are at greater risk for hypotension during high sensory level subarachnoid block than are women with a negative SST. A one liter colloid IV fluid preload may reduce their risk of hypotension.



This study has some important clinical lessons for us despite two important methodological problems.

Hypotension during regional anesthesia for cesarean section remains a problem. And hypotension is harder to treat in a pregnant woman due to the vulnerability of the fetus to hypoxia if the uterine arteries are constricted with a vasopressor. Knowing which women are at greatest risk for hypotension would be helpful. It appears that the Supine Stress Test (SST), detailed in this study, may be a useful clinical tool to identify these at risk women. The 90% positive predictive value says that if a patient has a positive SST there is a 90% chance she will experience hypotension during regional anesthesia for cesarean section. Unfortunately the opposite cannot be said. The negative predictive value was much lower, so if a SST is negative that doesn’t mean she won’t have hypotension. The SST can be performed with monitors we already have, in only a few minutes, with no additional preparation or training. It is worth a trip to the library or purchasing the full article online just to have a copy of the Supine Stress Test procedure.

Another useful aspect of this study which makes having the original article worthwhile is the questionnaire the investigators administered. I didn’t include it in the summary because the investigators didn’t say much about it. It appeared to be an “add on” that wasn’t well integrated into the study. Before the SST, the investigators asked each woman if they were able to sleep on their back during the last trimester or if they experienced nausea, vomiting, or faintness while lying on their back. Only 12 of 53 woman said they had experienced none of these symptoms while lying on their back and that they could sleep on their back. Of those 12 women only one became symptomatically hypotensive during her subarachnoid block. This information wasn’t analyzed and we certainly can’t draw any firm conclusions from it. I mention it simply to raise awareness in case similar questions show up in future studies. It may be that if women are asymptomatic while lying on their back at home and can sleep on their backs they are in a low risk group for hypotension. This is speculative but interesting. Left uterine displacement remains a standard of care during regional anesthesia in obstetrics.

Now to the bad news. The first problem this study suffers from is the implicit assumption that maternal hypotension during regional anesthesia is due to a relative hypovolemia which results in reduced venous return to the heart and, thus, a reduction in cardiac output. Scores of studies have searched for the volume of IV preload that fixes this maternal hypotension problem without finding it. Note that in this study 61% of women who received a 1 L colloid preload still experienced hypotension. One has to wonder when we will ask “what is causing the hypotension” rather than “how much volume does it take to fix the hypotension.” Maternal hypotension during epidural or subarachnoid block is, in my view, much more a function of reduced systemic vascular resistance than reduced venous return. The case for this position is laid out in the Obstetric Anesthesia Chapter1 of Nagelhout and Zaglaniczny’s textbook, “Nurse Anesthesia.” The good news is, while the “volume assumption” muddies the water it is unlikely to affect what is probably the most important information in the study. There is a relatively simple test (the Supine Stress Test) that is relatively good at predicting women at greatest risk for hypotension during regional anesthesia for cesarean section.

The second problem this study suffers from may have significantly impacted the results. Women who are in labor when their spinal or epidural block for cesarean section sets up experience less hypotension than women who are not in labor.2 If some of the study patients were in labor and others were not, the incidences and severity of hypotension reported in this study may be significantly biased. From the information presented we don’t know if this bias was a factor and, if it was, we don’t know in which direction it may have biased the results. I’m surprised the investigators overlooked this detail because they give all indications of being experts in obstetric anesthesia. It is certainly possible that they, in fact, did not overlook this important factor. They may have controlled for it and simply forgotten to write about it. In any case we don’t know if laboring or the lack of it affected the results of this study or not and that is a problem.


Michael Fiedler, PhD, CRNA


1.? Fiedler M. Obstetric anesthesia. In: Nagelhout JJ, Zaglaniczny KL, eds. Nurse Anesthesia. 3rd ed. St. Louis: Elsevier Saunders, 2005:1052-96.

2.?Brizgys RV, Dailey PA, Shnider SM et al. The incidence and neonatal effects of maternal hypotension during epidural anesthesia for cesarean section. Anesthesiology 1987;67:782-6.


The Positive Predictive Value of a test is the probability that an individual actually has what is being tested for when the test is positive.

The Negative Predictive Value of a test is the probability that an individual does not have what is being tested for when the test is negative.

© Copyright 2007 Anesthesia Abstracts · Volume 1 Number 5, August 31, 2007

Wong CA, McCarthy RJ, Fitzgerald PC, Raikoff K, Avram MJ

Gastric emptying of water in obese pregnant women at term

Anesth Analg 2007;105:751-755

Wong CA, McCarthy RJ, Fitzgerald PC, Raikoff K, Avram MJ



Purpose            The purpose of this study was to compare gastric emptying in obese term pregnant women after they ingested either 50 mL or 300 mL of water.

Background            Obesity is becoming more and more common in western pregnant women. Obese women are more likely than non-obese women to undergo a cesarean section. Studies in obese individuals have shown gastric emptying to be slower, unchanged, or faster than in non-obese individuals.

Fasting during labor and delivery or prior to elective cesarean section is uncomfortable, it may result in dehydration, and it increases patient anxiety. Studies in healthy nonpregnant adults have shown that consuming clear liquids up to two hours prior to induction of general anesthesia does not increase gastric volume compared to those who remained NPO overnight. The American Society of Anesthesiologists suggests that fasting guidelines for healthy women scheduled for elective cesarean section be the same as for nonpregnant women (allowing clear liquids up to two hours preoperatively).

Methodology            This prospective, single blind, randomized, crossover study included 10 term pregnant women whose Body Mass Index (BMI) was ≥ 35 kg/m2. Gastric size over time was determined by serial ultrasound exams, performed by the same ultrasonographer, using a standardized procedure and patient position. Gastric ultrasound was performed at baseline before water ingestion and then at 10 minute intervals for 60 minutes. The cross sectional area of the gastric antrum was used as a measure of gastric size. Each subject participated in both the 50 mL group and the 300 mL group. Participation in different groups was separated by at least 48 hours.

Result            Average gastric size was larger in the 300 mL group at 10 minutes but nearly the same at baseline, 20, 30, 40, 50, and 60 minutes. The rate of gastric emptying was no different in the 50 mL and 300 mL groups. Gastric volumes at 60 minutes were clinically and statistically no different than at baseline in either group.

Conclusion            In obese term pregnant women, gastric emptying was no different following ingestion of 50 mL or 300 mL water. This finding supports using the same fasting guidelines for healthy obese and non-obese term pregnant women prior to elective cesarean section.



This is really a much more complex topic than it may first seem. Gastric volume is an important factor, and we know that larger volumes of aspirate increase patient morbidity. The risk of regurgitation and pulmonary aspiration of gastric contents during induction of general anesthesia is, however, associated with many factors in addition to gastric volume. The good news from this study is that 60 minutes after drinking 300 mL of water (about 10 ounces) gastric volume had returned to baseline. This provides support for those of us who don’t see much reason to withhold reasonable amounts of water from pregnant women in an effort to reduce their aspiration risk if a surgical procedure is required. Sixty minutes after then drink a glass of water this study shows that their gastric volume will have returned to where it was before they drank it; and they’ll probably feel better. In fact, visual inspection of the figure in the study looks as if gastric volumes were back to baseline by 20 to 30 minutes. I find this information encouraging but it is not the whole story. It does not, for example, suggest that obese pregnant women are at the same risk of aspiration as nonpregnant women. (There is plenty of good evidence that their risk is increased.) Baseline gastric volume in an obese term pregnant woman may be markedly increased compared to nonpregnant women. All this study says is that 30 to 60 minutes after drinking 300 mL of water gastric volume is no greater than it was before they drank the water.

In my view, this study provides support for the practice of allowing obese term pregnant women to drink reasonable amounts of water up to an hour before induction of general anesthesia. It does not reduce my concern that term pregnant women have a full stomach and are at increased risk of aspirating gastric contents during anesthetic induction.


Michael Fiedler, PhD, CRNA

© Copyright 2007 Anesthesia Abstracts · Volume 1 Number 5, August 31, 2007


Hohlrieder M, Brimacombe J, Eschertzhuber S, Ulmer H, Keller C

A study of airway management using the ProSeal LMA laryngeal mask airway compared with the tracheal tube on postoperative analgesia requirements following gynaecological laparoscopic surgery

Anaesthesia 2007;62:913-918

Hohlrieder M, Brimacombe J, Eschertzhuber S, Ulmer H, Keller C


Purpose            The purpose of this study was to compare postoperative analgesic requirements in laparoscopic surgery patients whose airways were managed with a ProSeal laryngeal mask airway versus an endotracheal tube.

Background            Clinical and laboratory evidence indicates that stimulation from areas other than the surgical site influence postoperative pain sensation. The presence of a cuff in the hypopharynx is less stimulating than a cuff in the trachea. Indirect evidence from other studies suggest that an extraglottic airway may be associated with reduced postoperative analgesia requirements compared to patients whose airway was managed with an endotracheal tube.

Methodology            This prospective, randomized, double blind study included 100 consecutive ASA class I and II patients scheduled for elective gynecologic laparoscopic surgery. Their ages ranged from 18 years to 75 years. Exclusion criteria included known or predicted difficult airway, oropharyngeal pathology, chronic pain, and a body mass index > 35 kg/m2.

All patients were premedicated with 0.05 – 0.1 mg/kg oral midazolam. General anesthesia was induced with fentanyl 1-4 ?g/kg, propofol 2.5-3 mg/kg and rocuronium 0.6 mg/kg. Once twitches were abolished, a 7 mm endotracheal tube (ETT group) or a size 4 ProSeal laryngeal mask airway (LMA group) was inserted. Laryngoscopy was performed with a Macintosh blade for placement of both the ETT and the LMA. The ETT was placed in the usual manner. To place the LMA, a laryngoscopy was performed and a bougie was inserted into the esophagus. The drain tube of the ProSeal LMA was threaded over the bougie and guided into the oropharynx. The bougie was then removed. Both the ETT and the LMA were secured in the midline with tape. ETT cuff pressure was regulated at 20 cm H2O and LMA cuff pressure at 60 cm H2O throughout the case. Anesthesia was maintained with remifentanil 0.25-0.5 ?g/kg/min and propofol 75-125 ?g/kg/min. A 14 French gastric tube was placed in all patients. Mechanical ventilation was set identically in both groups. Forced air warming was used to maintain near normothermia. Diclofenac (Voltaren) 75 mg IV was administered during the case and 12 hours postoperatively. If necessary, rocuronium was antagonized at the end of the case with ≤ 2.5 mg neostigmine and atropine. At the completion of surgery the ETT or LMA was removed when the patient was breathing and opened their mouth to verbal command. In the PACU all patients received morphine IVPCA.

Result            The total dose of IVPCA morphine used at 2, 6, and 24 hours postoperatively was 30%, 31%, and 23% lower in the LMA group than the ETT group (P<0.05 at 2h and 6h). There was no difference in the incidence of sore throat, dysphonia, or dysphagia. No patient identified the airway as the source of their pain. Average pain scores were also lower in the LMA group at 2, 6, and 24 hours but the difference was neither statistically nor clinically significant.

Conclusion            The need for morphine postoperative analgesia is lower in gynecologic surgery patients whose airway was managed with a ProSeal LMA than in those managed with an ETT.



This is not a strong study methodologically or statistically. There was no evidence of a cause and effect relationship between the airway used (LMA or ETT) and the level of postoperative pain, only a weak association. One could easily make a case that I should not have included this study. I chose to include this study, however, because the concept is so interesting. We are just beginning to understand the complex generation / regulation of surgical pain. The pain perceived after a peripheral surgical wound is influenced by, among other things, local inflammation, tissue factors, excitatory and inhibitory neurons, and excitatory and inhibitory neurotransmitters. We know, for example, that blocking noxious stimuli with regional anesthesia intraoperatively reduces central sensitization resulting in less pain perception postoperatively. And we know that preventing inflammation with NSAIDS and/or steroids results in less pain than “treating” the inflammation after it occurs. Clearly, factors remote from the surgical site play a role in surgical site pain perception. These investigators hypothesized that noxious stimuli from other parts of the body might influence postoperative pain perception at the surgical site. They chose the stimulation of an ETT and an LMA to test their hypothesis and found just a little evidence to support their idea. I’m not ready to allow this study to influence my choice of an LMA or ETT for a given patient but it is certainly interesting to think about.

Michael Fiedler, PhD, CRNA

© Copyright 2007 Anesthesia Abstracts · Volume 1 Number 5, August 31, 2007


Hans P, Vanthuyne A, Dewandre PY, Brichant JF, Bonhomme V


Blood glucose concentration profile after 10 mg dexamethasone in non-diabetic and type 2 diabetic patients undergoing abdominal surgery

Br J Anaesth 2006;97:164-170

Hans P, Vanthuyne A, Dewandre PY, Brichant JF, Bonhomme V


Purpose            The purpose of this study was to compare intraoperative blood glucose concentrations in non-diabetic and type 2 diabetic patients after dexamethasone 10 mg IV. A secondary aim was to measure the associations between 1) Hemoglobin A1c (HbA1c) and the percent increase in blood glucose and 2) Body Mass Index (BMI) and the percent increase in blood glucose.

Background            Dexamethasone contributes to the prevention of Postoperative Nausea and Vomiting (PONV), especially when combined with other antiemetic drugs. Studies of dexamethasone and PONV have used doses of 8 mg and, in some studies, even higher doses. Dexamethasone has also been shown to increase blood glucose concentrations during surgery. This is true even in non-diabetic patients. Elevated blood glucose increases the risk of a poor outcome in patients at risk for central nervous system ischemia.

Methodology            This prospective study included 63 consecutive ASA class II patients undergoing early morning elective abdominal surgery. Group ND was non-diabetic and included 32 patients. Group D had type 2 diabetes treated with oral anti-diabetic agents and included 31 patients. No patient received dexamethasone, insulin, or an anti-diabetic drug before induction of general anesthesia.

All patients were premedicated with 0.5 mg alprazolam (Xanax) and 0.5 mg atropine orally one hour before induction. Anesthesia was induced with 0.15 ?g/kg sufentanil, 2 mg/kg propofol, and 0.15 mg/kg cis-atracurium. Anesthesia was maintained with sevoflurane and 50% nitrous oxide. Dexamethasone 10 mg was given to all patients during induction. Clonidine 300 ?g was infused over 15 minutes at the start of surgery.

Blood glucose was measured by finger stick immediately before dexamethasone administration. This was Time Zero (T0). Blood glucose was also measured at T1, T2, T3, and T4 which were 60, 120, 180, and 240 minutes after T0 respectively.

Result            At T0, immediately before dexamethasone administration, blood glucose averaged 127 mg/dl (sd 20 mg/dl) in the Diabetic group and 105 mg/dl (sd 11 mg/dl) in the Non-diabetic group. The maximum glucose concentration was 162 mg/dl (range 121 – 232 mg/dl) in group D. This was a 29% increase from baseline. The maximum glucose concentration was 142 mg/dl (range 104-180 mg/dl) in group ND. This was a 35% increase from baseline. Average glucose peaked at T2, 120 minutes after dexamethasone administration. Average glucose was higher in group D at all time points. Patients in group D were older than patients in group ND.

Considering both group ND and group D patients together, the maximum blood glucose concentration correlated linearly with HbA1c (R2=0.26, P<0.01) and BMI (R2=0.21, P<0.01).

One patient in group ND was diagnosed with diabetes during the study. This patient nonetheless remained in, and was analyzed as part of, the ND group.

Conclusion            Following dexamethasone 10 mg IV glucose levels were significantly higher in type 2 diabetic than non-diabetic patients for at least 4 hours. The increase in glucose levels peaked at 2 hours after dexamethasone administration. The increases in blood glucose seen in this study may not be clinically significant under most circumstances. Under some circumstances, however, glucose levels greater than 140 mg/dl may be associated with increased risk of an adverse outcome. (Examples might include cardiac surgery and stroke patients.)


Dexamethasone is now understood to be quite useful both in helping to prevent PONV and in reducing the inflammatory component of acute surgical pain. It is being used more and more as part of a multimodal PONV and pain reduction strategy. Previous studies have shown that dexamethasone 10 mg and less has few, if any, adverse effects that would preclude its use for these purposes. What has not been thoroughly investigated is the effect of dexamethasone on blood glucose levels. Glucose levels are important for a number of reasons that are important to anesthesia, including diabetics and patients experiencing or at risk for CNS ischemia. In my mind, the best thing this study does for us is give us an idea what glucose levels are likely to do during surgery and general anesthesia in type 2 diabetic and non-diabetic patients who have received dexamethasone. The correlations between HbA1c or BMI and blood glucose increases are small but very real. So, as HbA1c or BMI goes up, blood glucose will increase more. But it takes a rather large increase in HbA1c or BMI for a modest increase in blood glucose. Unfortunately, what this study doesn’t tell us is if the increases in blood glucose after dexamethasone are any greater than the increase in blood glucose seen when dexamethasone is not given.

The investigators correctly point out a significant limitation of this study. Blood glucose levels have been shown to increase during general anesthesia and surgery when dexamethasone has not been given. The increase in glucose during surgery is attributed to the stress response and is not seen during regional anesthesia. Because there was no control group in this study, we don’t know how much the glucose levels would have increased without dexamethasone. Thus, we certainly can’t attribute all the increase in blood glucose levels to the dexamethasone administration. Hindsight is great, but it sure would have been nice to have had one more group that didn’t get dexamethasone so we’d know what the glucose levels would have been without it. The sufentanil administered equates to about two mL fentanyl so it probably did not impact the stress response much over the four hour study period.

I have no idea why the investigators left the diabetic patient in the non-diabetic group except that doing so was the statistically “pure” way to conduct the study. Fortunately, this choice does not appear to have affected the study results in a way that is clinically important.

Lastly, as a side note, this study reported blood glucose levels in mmol/L (millimoles per liter). In the USA we are used to seeing glucose expressed in mg/dl (milligrams per tenth of a liter). The conversion between mmol/L and mg/dl is specific to the substance being measured, in this case glucose, because the molecular weight of the substance must be considered in the conversion. To convert between mmol/L of glucose and mg/dl simply multiply the mmol/L reported by 18. This conversion should come in handy if you read the original article.


Michael Fiedler, PhD, CRNA


D'Alecy LG, Lundy EF, Barton KJ, Zelenock GB. Dextrose containing intravenous fluid impairs outcome and increases death after eight minutes of cardiac arrest and resuscitation in dogs. Surgery 1986;100:505-11.

© Copyright 2007 Anesthesia Abstracts · Volume 1 Number 5, August 31, 2007

Sacan O, White P, Tufanogullari B, Klein K


Sugammadex reversal of rocuronium-induced neuromuscular blockade:  a comparison with neostigmine-glycopyrrolate and edrophonium-atropine 

Anesth Analg 2007;104:569-574

Sacan O, White P, Tufanogullari B, Klein K



Purpose            The purpose of this clinical research study was to determine if sugammadex, a new investigational g cyclodextrin compound recently developed for the reversal of certain neuromuscular blocking agents, would demonstrate a more rapid reversal when a profound rocuronium-induced blockade is present compared to the traditional cholinesterase inhibitors neostigmine and edrophonium. Additionally, intra-operative and post-operative side affects of the new compound were assessed and compared with any side affects observed from the use of neostigmine or edrophonium. Some basic assumptions of this study include the fact that sugammadex is a chelating or encapsulating agent only to be used for reversal of steroid based nondepolarizing neuromuscular blocking drugs specifically rocuronium and possibly vecuronium. Other basic assumptions include the fact that side affects typically seen from the cholinesterase inhibitors not only include those from the inhibitory affects themselves, but also include side affects from the muscarinic antagonistic affects of glycopyrrolate and atropine. Since Sugammadex is administered in isolation and not in combination form, the muscarinic antagonistic side affects such as tachycardia, would be potentially minimal.

Background            The traditional cholinesterase inhibitors, most notably neostigmine and edrophonium, are used quite frequently, and are considered to be safe and efficacious. As with many of the traditional and older anesthetics there is always room for attempting elimination of side affects. Efforts to minimize untoward affects do exist, and there are still key limitations to their use. Key issues that have not been solved include the fact that these reversal agents are not safe nor appropriate to use when “deep” residual blockade is still present, and that they are known to cause, either alone and/or when used with the muscarinic antagonists, tachycardia, bradychardia, dry mouth and nausea and vomiting. Organon USA, Inc, (Roseland, NJ) has developed a new class of drug to reverse rocuronium induced neuromuscular blockade. This new class of drug is a g cyclodextrin, and previous clinical trials have shown that at certain doses, the decreased median recovery time to achieve a train-of-four ratio response of 0.9 decreased from 21 minutes (with the traditional agents) to 1.1-1.3 minutes with the use of Sugammadex. The cyclodextrins form tightly bound 1:1 complexes with aminosteroid-based muscle relaxants, specifically rocuronium, and act by encapsulating the rocuronium. When sugammadex encapsulates, this initially increases the plasma concentration of the muscle relaxant, and therefore there are minimal relaxant molecules at the neuromuscular junction. This results in a rapid reversal action of the residual neuromuscular blockade from the rocuronium.

Methodology            This research was conducted as an open, parallel study design. IRB approval was obtained and 60 patients were enrolled (determined via a statistical power analysis in order to detect a difference in the return to TOF ratio of 0.9 within fives minutes after receiving sugammadex, compared with 25% or less of those who received either of the cholinesterase inhibitors), 20 in each of three different groups. The three groups received one of the following reversal agents after a very specific and standardized anesthetic protocol:  Group 1 sugammadex, Group 2 neostigmine and glycopyrolate, and Group 3 edrophonium and atropine. The patients who were to undergo elective surgical procedures requiring tracheal intubation were enrolled, and if any patient preferred not to receive the investigational drug (sugammadex) they were randomly assigned to one of the anticholinesterase groups. The reversal drugs were given at least 15 minutes after the last dose of rocuronium. The maintenance anesthetic agents used were continued for 30 minutes after the reversal agents were given, as well as each patient’s neuromuscular assessment (monitoring done for the 30 minutes). The neuromuscular assessment of the adductor pollicis muscle was monitored using an acceleromyograph that was calibrated, stabilized for each patient, and transferred in real-time to a laptop computer in the operating room. Each patient’s hemodynamic status (heart rate) was recorded at predetermined intervals throughout the reversal and emergence time frame. Clinical signs of anesthetic recovery were also monitored and documented in a standardized mode. Adverse events were recorded by a blinded observer in the operating suite and upon discharge from the recovery room. All appropriate demographic data, specifics of neuromuscular function, and hemodynamic differences were recorded, evaluated, and compared across the three groups.

Result            The predetermined sample size needed to detect significant differences across groups was achieved; each group consisted of 20 patients. There were no significant demographic differences noted between groups that could account for the results obtained. In addition, the total length of anesthesia time and the time to administration of the reversal drug after the last dose of rocuronium were not different across the three groups. The height of the first twitch at the time that the reversal agents were given was similar in all three groups. Of statistical significance though was the time to achieve a train-of-four ratio (TOF) of 0.7, 0.8 and 0.9 and this was prolonged in the edrophonium and neostigmine groups. All of the sugammadex patients achieved a TOF ratio of 0.9 in less than 5 minutes compared with none and one in the edrophonium and neostigmine groups, respectively. The patients who received neostigmine and glycopyrrolate demonstrated significantly higher heart rates values at 2 and 5 minutes after the reversal drugs were given. Sugammadex was shown to cause less of a dry mouth complaint from patients in the recovery room. There were no significant differences seen across the three groups in terms of nausea, vomiting, drowsiness or dizziness.

Conclusion            This studied demonstrated that sugammadex was associated with a more rapid and complete reversal from what is termed a moderately profound neuromuscular block, when rocuronium, an aminosteroid neuromuscular blocking agent, is used. Using sugammadex alone instead of having to combine it with an anticholinergic (as the traditional reversal agents warrant) would clearly eliminate any untoward affects from the use of an anticholinergic, such as tachycardia and using it alone, due to its pharmacodynamics, does not cause untoward affects such as bradycardia. This study however was not without limitations. The sample size may not have been large enough to detect any other untoward affects, for example post operative nausea and vomiting, typically seen when anticholinesterases are used. Additionally the types of surgery conducted, i.e. urology procedures, have a low incidence of the specific ill-affects of the antichoinesterases as compared to females undergoing gynecologic procedures. This was an open-parallel study design and one can always make the argument that the gold standard in clinical trials is viewed as a prospective randomized double blinded design.



This was a very rigorous, well conducted, and easy to follow study which appears to have major ‘positive’ implications in the future practice of anesthesia. There are always going to exist times when administration of a depolarizing muscle relaxant is contraindicated. Even when all evidence appears that the patient is “intubate-able” there will occasionally be times when, after paralysis with a nondepolarizing relaxant, the patient cannot be intubated. And, worse yet, some of these times mask ventilation will also be difficult or impossible. Being able to rapidly reverse a non-depolarizing muscle relaxant when an airway is critically threatened is a phenomenally positive capability.

As the authors do clearly point out, the pharamaco-economics of this new drug will certainly play a role in its acceptance for use by all providers. It is always perplexing, though, that we have been trained to consider weighing the cost of a drug against the benefits of what we are trying to achieve, rather than thinking only of the best possible patient care. Sugammadex may, however, be so unique and effective that we have the rare freedom to consider only its advantages to our patients.


Mary A. Golinski, PhD, CRNA




© Copyright 2007 Anesthesia Abstracts · Volume 1 Number 5, August 31, 2007

Ray DC, McKeown DW


Effect of induction agent on vasopresor and steroid use, and outcome in patients with septic shock

Crit Care 2007;11:1-8

Ray DC, McKeown DW



Purpose            Patients with sepsis and septic shock exhibit hypoxemia, hypotension, volume depletion, renal and other organ impairment. Administration of an induction agent, for surgery or placement of an endotracheal tube for those needing mechanical ventilation, can be detrimental to the patient and often times cause worsening of hemodynamic compromise. Etomidate is known to minimally affect cardiovascular instability however it does cause adrenal suppression. It has been postulated that adrenal suppression and the potential reduction in intrinsic steroid synthesis could increase the need for vasopressors and steroid therapy in those exhibiting septic shock if given etomidate for the induction of anesthesia. The purpose of this study was to assess if an association exists between induction agent used, and the need for vasopressors, inotropes, and steroids in those with septic shock. Additionally, the researchers studied the association between induction agent and outcomes of patients who suffered from septic shock.

Background            It has not been determined which induction agent used for airway establishment with endotracheal tube placement, is the “best” for those with sepsis and/or septic shock. Current choices include, but are not limited to, propofol, thiopental, etomidate, midazolam, and ketamine. It is well known that propofol causes the most cardiovascular depression relative to the other agents, and that etomidate causes the least cardiovascular depression compared to propofol or thiopental. Past research has demonstrated that there is a degree of adrenal function suppression exhibited when etomidate is used; it is related to blockade of 11 B-hydroxylase. It remains controversial as to how long this adrenal suppression lasts; some studies suggest it may last up to 72 hours. It is not clear if there are true clinical consequences owed to the adrenal suppression, particularly in those with septic shock and there is very little evidence that this adrenal suppression is related to patient outcome. Previous studies have suggested that etomidate may be associated with an increased mortality in those with sepsis and that corticosteroids used for patients with sepsis may improve outcomes.

Methodology            A retrospective chart review and analysis was conducted for patients previously admitted (over a 40 month period) to an adult intensive care unit in a major teaching hospital and tertiary referral center with critical illnesses. Exclusion criteria were upheld for those post- cardiac surgery, those with uncomplicated cardiology problems, and those with isolated head injuries. Due to the nature of the retrospective analysis, informed consent was not required by the local ethics committee. The Scottish Intensive Care Society WardWatcher ™ database was used to record details of each patient, including reason for admission, diagnosis, patient outcomes, and predicted outcomes. Of patients admitted in this 40- month period (n = 2054) 242 patients were diagnosed with septic shock, and 208 of the 242 required tracheal intubation and mechanical ventilation. Complete information was available to the researchers for 159 patients  (who received an induction agent and subsequent intubation) to assess:  patient demographics, source of sepsis, admission and outcomes details, the sequential organ failure assessment score within the first seven days of their stay in the ICU. Additionally documented: which induction agent was administered, the dose and duration of vasopressor or inotropic support at the time of induction, and doses of steroid administration at the time of induction. The researchers noted whether any significant cardiovascular compromise was present at the time of the induction. Statistical analysis was used to test and determine if significant differences were seen regarding vasopressor use, steroid requirements, and patient outcomes based on which induction agent was used.

Result            There were no differences noted in the demographic data for all patients included in the study (n =159). The agents (the number of patients) used for the induction of anesthesia were etomidate (74), propofol (25), thiopental (26), midazolam (14), ketamine (1),  fentanyl (1), and inhalation induction (2). Sixteen patients (16) did not have an induction agent administered at all (for example, endotracheal tube placement was during CPR or awake fiberoptic intubation); 75 of the total sample of patients were intubated in areas outside of the ICU, mostly in the operating room or emergency room. The severity of illness details and the outcomes for each induction agent used (for each patient) did not show statistical significance. For example, the age of the patient, their APACHE II score, their predicted mortality and hospital mortality and their organ failure assessment scores did not differ between those who received different induction agents. All 159 patients received vasoactive drug infusions for blood pressure support and the mean numbers of infusions per patient did not differ depending on induction agents. The choice of induction agent used did not correlate with the timing of initiating noradrenaline infusions (it didn’t matter which induction agent was used as to the timing of a noradrenaline infusion ) nor did it influence additional and subsequent steroid administration (eighty-seven patients required steroid therapy for blood pressure support). Statistical significance was not seen between those given etomidate who received steroids and died, compared with those who were given etomidate, did NOT receive steroids, and who died. For those who received etomidate, vasopressor use was less frequently required, but this was not statistically significant.

Conclusion            In general, for those patients whose cases were retrospectively reviewed in this research, it was found that vasopressor use and the dosage of vasopressors were not influenced by the anesthetic induction agent. Additionally, steroid use for those with vasopressor-resistant hypotension was not influenced by the induction agent used. When the patients were at considerably high risk due to pre-existing factors, their outcomes were not influenced by the administration of etomidate. Outcomes were not improved either, if steroids were administered to those patients given etomidate.



It remains extremely challenging to receive a patient with sepsis or septic shock with extreme hemodynamic compromise, necessitating an anesthetic, either for placement of an endotracheal tube or for pending surgery. Simply organizing the reasons for the various and multiple vaso-active infusions as well as considering their actions on the cardiovascular, pulmonary, neurologic and renal systems takes a great deal of focus and understanding of the disease process as well as the pharmacokinetic models of the agents. The multi-system organ involvement for those with sepsis and/or septic shock can be overwhelming and determining which anesthetic agents will either promote cardiovascular stability or prevent a worsening of hypotension and other negative clinical physiological sequela is a complex process. Understanding that retrospective chart reviews do impose certain limitations, this study offers the provider evidenced based data and the ability to weigh the risk benefit ratio when faced with such a scenario.


Mary A. Golinski, PhD, CRNA




© Copyright 2007 Anesthesia Abstracts · Volume 1 Number 5, August 31, 2007

Regional Anesthesia

Wang LZ, Zhang YF, Tang BL, Yao KZ


Effects of intrathecal and IV small-dose sufentanil on the median effective dose of intrathecal bupivacaine for caesarean section

Br J Anaesth 2007;98:792-796

Wang LZ, Zhang YF, Tang BL, Yao KZ



Purpose            The purpose of this study was to determine the location of action of a 2.5 ?g subarachnoid dose of sufentanil.

Background            Subarachnoid block is commonly used for cesarean section. Adding fentanyl class opioids to the local anesthetic improves the quality of such blocks. Sufentanil is considerably more lipid soluble than fentanyl. Sufentanil provides longer lasting and more potent analgesia than fentanyl when added to local anesthetic for subarachnoid block. There is controversy about the site of action of sufentanil when it is added to local anesthetic for a subarachnoid block. Some information indicates that subarachnoid sufentanil produces analgesia in the spinal cord. Other information indicates that sufentanil is absorbed systemically and acts in the brain, much like an intravenous dose.

Methodology            This prospective, double-blind, randomized study included 90 ASA class I and II term pregnant women scheduled for elective cesarean section. Women with pregnancy induced hypertension, fetal heart rate abnormalities, or who weighed >100 kg were excluded. The women were divided into three groups. Group C (control) received no sufentanil, group IVS received 2.5 ?g sufentanil intravenously (IV), and group ITS received 2.5 ?g sufentanil in their subarachnoid block mixed with local anesthetic.

All women were premedicated with metoclopramide and ranitidine. All received nasal cannula oxygen and 500 mL crystalloid IV fluid. Next, a combined spinal-epidural (CSE) block was performed at L3-4 with a 16 gauge Touhy and a 26 gauge pencil-point spinal needle. A subarachnoid block was established with 0.5% bupivacaine with dextrose and with or without sufentanil 2.5 ?g, depending upon the study group. All blocks were performed in the lateral position with the opening of the spinal needle facing cephalad. Immediately after the block, an epidural catheter was placed through the Touhy needle. Patients were then positioned supine with left uterine displacement. Simultaneously, 2.5 ?g sufentanil was injected IV in the IVS group or an equal volume of saline was injected IV in groups C and ITS.

The first patient in each group received 9 mg of bupivacaine. Subsequent patients received 1 mg more if the previous patient’s block was inadequate or 1 mg less if the previous patient’s block was adequate. An inadequate sensory block was below T-6 or required supplementation for patient comfort. An adequate sensory block was T-6 or above and did not require supplementation for patient comfort. When supplementation was required, 2% lidocaine was injected through the epidural catheter. Using this method, the ED50 (effective dose in 50% of the population tested) and 95% confidence level for bupivacaine with or without subarachnoid or IV sufentanil was determined.

Result            The ED50 for bupivacaine was 6.3 mg in group C, 5.2 mg in group IVS, and 3.0 mg in group ITS (P<0.0005 ITS group compared to C and IVS groups). Seven patients (23%) in the ITS group reported itching compared to no patients in the other two groups (P<0.01). Nausea, vomiting, and bradycardia were uncommon and not different between groups. The incidence of hypotension was not statistically significantly different between groups but there was a clinically noticeable trend toward less hypotension in the ITS group than in the IVS or C groups. The average Apgar score was 9 or greater in all three groups at both 1 and 5 minutes.

Conclusion            Subarachnoid, but not IV, sufentanil 2.5 ?g had a dose-sparing effect on the dose of hyperbaric bupivacaine required for a successful block for cesarean section. Subarachnoid sufentanil 2.5 ?g enhanced the potency of subarachnoid block anesthesia. This suggests that 2.5 ?g sufentanil injected in the subarachnoid space works in the spinal cord.



This is a simple and well executed study that demonstrates convincingly that 2.5 ?g sufentanil enhances the anesthesia of a bupivacaine subarachnoid block by working in the spinal cord. Adding this small dose of sufentanil reduced the dose of bupivacaine needed to produce a T-6 block by about half.

It is the actual dose of bupivacaine used for these blocks about which I would most like to comment. First, while the ED50 of bupivacaine for a T-6 sensory block was 3 mg in the subarachnoid sufentanil group (ITS group) I don’t think anyone would suggest using 3 mg clinically. The ED50 is simply a concept used to compare potency. For inhalation agents we call the ED50 “MAC.” And just like we don’t plan our inhalation anesthetics to be sufficient in 50% of patients, we don’t plan our spinals to work in 50% of patient either. ED50 was used by these researchers simply as a convenient way to compare the dose needed by different groups. Second, I usually dose C-section patient sitting up. Because bupivacaine starts to set up so quickly, I use a larger dose of bupivacaine in sitting patients than I do in side lying patients. All these women were dosed in the lateral position minimizing the dose that would be required in any case. Despite all this, I was surprised by the data presented in the dosing figures in the original article. No women in the subarachnoid sufentanil group (ITS) who received 5 mg bupivacaine or more had a block that needed supplementation. This is a much lower dose than I would have predicted and I can’t explain it. The study was performed in China and perhaps western women would require a larger dose for physiologic or cultural reasons.

I have been adding up to 25 ?g fentanyl to my subarachnoid blocks (unless contraindicated) for almost two decades. I’ll now need to consider using sufentanil. And while I don’t expect to be doing C-sections with 3 mg or even 5 mg bupivacaine spinals, I will be thinking about lower doses than I have used in the past … when 2.5 ?g sufentanil is added.


Michael Fiedler, PhD, CRNA




© Copyright 2007 Anesthesia Abstracts · Volume 1 Number 5, August 31, 2007

Mitra R, Fleischmann K



Management of the sheared epidural catheter: is surgical extraction really necessary?

J Clin Anesth 2007;19:310-314

Mitra R, Fleischmann K




Purpose            The purpose of this article was to review the available scientific evidence regarding management of a fragment of epidural catheter broken off within a patient.

Background            Fragments of epidural catheters occasionally break off in patients due to shearing or breaking. Breaking is usually the result of excessive force being used to pull on an immovable catheter. Epidural catheters may become immovable due to impingement of the catheter by anatomical structures, by tangling of the catheter with anatomic structures, or by knotting of the epidural catheter around itself. Epidural catheters threaded more than 5 cm into the epidural space have been shown to be more likely to knot on themselves. No formalized algorithms for management of a retained epidural catheter fragment are available. Many case reports describe clinical experiences with retained epidural catheters. A few studies examine various aspects of the issue, such as the force required to result in epidural catheter breakage.

Methodology            The authors performed a literature search of eight scientific databases. The following search terms were used: anesthesia, epidural, catheterization, retained, sheared. The authors organized and analyzed the information located in the search.

Result            The following information is divided into three subdivisions: Catheter Removal, Breaking, and Management of a Retained Catheter Fragment.

Catheter Removal

In general, less pulling force was required to remove an epidural catheter when the patient was positioned laterally than when sitting. Specifically, less force was usually required when attempting to remove the epidural catheter with the patient in the same position (lateral or sitting) as during epidural insertion. When difficulty removing the epidural catheter was encountered despite optimal positioning, extreme flexion has been recommended. The use of a Wilson Convex Surgical Frame has been reported in the successful removal of a previously immovable epidural catheter. When efforts to remove an epidural catheter fail, ceasing traction on the catheter for a period of time (even several hours) and trying again later has been reported to result in “easy” removal of a catheter.


A number of studies have examined the conditions under which epidural catheters break. One such study compared the stretching force required to cause an epidural catheter to break. The study compared the following epidural catheters:

  • 19-gauge Flex Tip Plus Arrow Catheter
  • 19-gauge Perifix Catheter
  • Perisafe Catheter
  • Portex Catheter

The authors reported that the Arrow Catheter stretched more and broke with less force than other catheters.

Another study examined the amount of stretching and force required to break catheters made of the following materials:

  • Polyurethane
  • Radiopaque
  • Clear nylon

The authors reported that the polyurethane catheters did not break within the limits of their testing. The radiopaque catheters were the most elastic.

A study examining the force required to break an epidural catheter compared catheters from the following manufacturers:

  • Abbott (nylon)
  • Baxter (nylon)
  • Becton Dickinson (nylon)
  • Burron (polyamide)
  • Concord-Portex (nylon)
  • Kendall (nylon)

The authors reported that Abbott catheters required the most force to break and Baxter catheters required the least force to break.

Management of a Retained Catheter Fragment

There are few case reports of the management of retained epidural catheter fragments. Received knowledge considers properly placed epidural catheters to be sterile, inert, and unlikely to result in neurologic complications. Based upon this rational, in adults many consider it safe to leave retained epidural catheter fragments in situ in the absence of neurologic complications. In contrast to this conventional wisdom, published reports suggest surgical removal of retained catheters.

In one such report, a 64 year old man had an epidural catheter fragment left in place for 18 months. At that time, he developed pain and weakness. A posterior epidural cystic mass was detected. The mass was surgically removed and, while the complication was attributed to the retained epidural catheter fragment, no catheter fragment was removed with the mass. In another case report a 34 year old woman developed L-3 radiculopathy seven months after an epidural catheter fragment was left in place. Imaging showed a coiled epidural catheter in direct contact with an L-3 nerve root. Surgical decompression resulted in complete relief.

When a fragment of epidural catheter breaks off in a patient the initial step is to identify where it lies. If it lies outside the spinal canal local anesthesia and a small incision may result in easy removal. If it lies within the spinal canal the risks and benefits of removing the fragment are harder to evaluate. While MRI has most often been used to localize catheter fragments, CT has been shown to produce clearer images of epidural catheters. Available literature does not support a significant risk of subsequent infection due to a retained epidural catheter fragment.

Conclusion            The substance of the authors recommendations include the following:

  1. Use gentle, continuous force when removing epidural catheters
  2. If excessive resistance is encountered stop, reposition the patient, and try again
    1. Place patient in same position as during epidural insertion
    2. Place patient in lateral position
    3. Place patient in extreme flexion
    4. Place patient in extreme extension
  3. If positioning does not allow easy catheter removal consider a CT to identify a reason for catheter immobility
  4. Stop pulling if the catheter begins to stretch
  5. If stretching occurs, wait and try again later
  6. If the catheter breaks consider leaving it in place in adults
  7. If neuropathy exists or a retained fragment lies within the spinal canal consult a neurosurgeon



The topic of managing epidural catheters that are difficult to remove, or catheter fragments broken off in patients, is one where there is far more clinical experience than research based knowledge. This article does a good job of summarizing the experience and research that has been published to date.

Over the years I have encountered a number of epidural catheters that were difficult or (temporarily) impossible to remove. The authors’ recommendations for approaching a difficult to remove catheter are consistent with what I understand. In my experience, if it is not possible to remove the epidural catheter without excessive force any position is better to try than the one the patient is in. Also, when a decision is made to stop pulling and try again later, I’m not sure the length of time one waits is as important as allowing some time for the patient to “wiggle around” on their own. I recall a time three different anesthesia providers attempted to remove an epidural catheter with the patient in various positions. We gave up and ordered a plain radiograph to try to see where the catheter was hung up. When the x-ray came back there was no epidural catheter visible on the film. I returned to the bedside only to find the entire epidural catheter in the sheets with the patient. Apparently she had moved in just the right way and the process of pushing the x-ray film under her had pulled the catheter out without either she or the x-ray tech even noticing. In my mind, patience has a large role to play in removing a stubborn epidural catheter.

Once broken, leaving an epidural catheter fragment in a patient is controversial. There is merit in the authors’ recommendation that the position of the catheter fragment within or outside the spinal canal be established. This is valuable information when deciding how to manage the retained catheter fragment. Their recommendations for conditions under which a neurosurgical consultation should be sought also make a lot of sense. To their recommendations please allow me to add one more. Since there is no clear consensus on whether an asymptomatic catheter fragment should be left in situ or retrieved surgically a judgment call will be required. And since anesthesia is by nature a consultation service, the patient’s primary care provider should be briefed about the retained epidural catheter. I would expect that the primary care physician (probably most often a surgeon) would want to make the judgment call to leave or retrieve a catheter fragment. I believe my job in that situation is to describe the specifics of the retained epidural catheter in their patient and provide information about the complication in general and likely sequela.


Michael Fiedler, PhD, CRNA





© Copyright 2007 Anesthesia Abstracts · Volume 1 Number 5, August 31, 2007