View from an Anesthesiologist

The question of who may administer endoscopic sedation has burned for over a decade, but recent developments have fanned the flames again. The approval by FDA of the SedaSys system and its subsequent withdrawal from the market has increased the heat, but not the light, surrounding this controversy. I am hopeful that this pro-con discussion can facilitate a cooling of passions. There is not a clear answer to “what is best for the patient.” If more patients undergo screening colonoscopy because they can get anesthesiologist-administered propofol, this is good. If the cost of anesthesiologist-directed propofol causes patients to avoid colonoscopy due to increased co-payments, this is bad. Given the low rate of complications of endoscopic sedation and the lack of an objective standard for evaluating the quality of endoscopic sedation, it is unlikely that we will have high-level evidence that answers the question of “what is safe”, much less “what is good”. So, let us turn to what is possible.

As an anesthesiologist who is actively involved in the development of automated delivery of propofol, I am often confronted by colleagues who tell me that I’m trying to put them out of a job. “It is not a job that you have,” I tell them, “and it is not a job that you want.” According to the Department of Labor, there are about 30,000 anesthesiologists employed in the U.S. The number of colonoscopies performed annually in the U.S. may exceed 30 million. If 10 percent of U.S. anesthesiologists staff 50 cases a day, five days a week, 50 weeks a year, we could achieve 100 percent coverage. I’ll leave the discussion regarding the cost of this to others, but it is unrealistic to propose 100 percent of cases be performed under a traditional anesthesia care model. Therefore, we must look to other models.

A number of alternative models have been proposed that shift administration of propofol from anesthesia providers to gastroenterologists and nurses who are not “trained in the administration of general anesthesia.” These models presume that a gastroenterologist can safely administer propofol to a subset of patients. While this approach has been validated, it has not been widely embraced. I’m not questioning John Vargo’s or Doug Rex’s ability to deliver propofol, but there are factors that will hinder wide adoption of propofol administration by nonanesthesiologists.

First, the gastroenterologist providing propofol sedation will almost certainly do so under a protocol crafted by a committee. Most committees will stick to that which can be easily measured — milligrams of propofol, possibly infusion rates. Since there is no readily available technology for real-time measurement of propofol blood levels, nor are there target-controlled infusions that could estimate these levels available in the United States, effect site concentrations will not be part of the protocol. Using drug administration, rather than effect site concentration, increases the uncertainty in propofol dosing by a factor of two.2

Data from my research group suggests a range of effect site levels required for endoscopy varies by a factor of four.3 These uncertainties cascade, so even if the committee would agree that the propofol limit should be based on some statistical measure of anesthesiologist practice, this number would be far from the target in many patients. A more likely outcome would be that the committee would choose to limit the dose based on a feared complication such as airway obstruction. Our data from studies of drug-induced sleep endoscopy suggests that the probability of airway collapse at effect site concentrations below 2 µg/ml is negligible.4 The problem is that this is 50 percent of the median effect site concentration in our endoscopy cohort, and given the skewed nature of the distribution, it is likely that a policy that tried to avoid obstruction would fail to achieve unresponsiveness in many patients and would fail to produce amnesia in some patients. Anesthesiologists deal with this uncertainty by adjusting the dose based on observation of the response, and gastroenterologists could likely learn this skill but will be restricted by hospital policy from exceeding this fixed dose. At my institution, midazolam dose in the endoscopy unitis restricted to a maximum of 10 mg by a policy that has not been revised since 1986, and there is no reason to suspect that propofol will be any different.

A second factor is the need to have anesthesiologist backup for a subset of patients. Again, a committee will decide the criteria comorbidities, history of difficulty with sedation, current drug use etc. Can we predict those patients who will fail to achieve sedation under our policy? Our data suggests not. When assessing the actions of experienced certified registered nurse anesthetist providers, we found that only 36 percent of the variability in propofol administration was programmed prior to initiation of endoscopy. We react far better than we predict. Can we predict which patients will experience airway obstruction? In our experience with drug-induced sleep endoscopy in patients with severe obstructive sleep apnea, we could not predict whether airway maneuvers would be required. If a model that utilizes anesthesiologists for patients predicted to exceed the protocol dose or predicted to have complications (such as obstruction) allocates too many patients to the wrong pathway, it will fail. If there is a job I would like less than staffing 50 colonoscopies every day of my life, it is one where I am constantly running from room to room trying to deal with inadequate sedation and respiratory arrests (with the ratio determined by a committee over which I had no influence). Given these issues, it is likely that many centers will find that models that attempt to cherry pick easy cases for gastroenterologist-directed propofol with an anesthesiologist available for the hard ones will find staff retention problematic.

Is there an alternative? My group has described approaches to pharmacokinetic control of propofol that simplify the task of the anesthesiologist, permitting a different staffing model, perhaps 1:35,6. Such a model would likely address both the manpower and cost issues associated with endoscopic sedation. Of course, this will require buy-in from third-party payors, professional organizations and hospital committees, assuming we can gain FDA approval and attract investors after what happened with SedaSys. Maybe our little brush fire isn’t so bad after all.

Dr. Mandel serves on the board of directors for the Society for Technology in Anesthesia.


1. Irwin RS, Madison JM. The diagnosis and treatment of cough. N Engl J Med. 2000;343(23):1715-21.

2. Hu C, Horstman DJ, Shafer SL. Variability of target-controlled infusion is less than the variability after bolus injection. Anesthesiology 2005;102:639–45.

3. Mandel JE, Sarraf E. The variability of response to propofol is reduced when a clinical observation is incorporated in the control: a simulation study. Anesth Analg 2012;114:1221–9.

4. Atkins JH, Mandel JE,Rosanova G. Safety and efficacy of drug-induced sleep endoscopy using a probability ramp propofol infusion system in patients with severe obstructive sleep apnea. Anesth Analg 2014;119:805–10.

5. Mandel JE, Lichtenstein GR, Metz DC, Ginsberg GG, Kochman ML. A prospective, randomized, comparative trial evaluating respiratory depression during patient-controlled versus anesthesiologist-administered propofol-remifentanil sedation for elective colonoscopy. Gastrointest Endosc. 2010;72:112-7.

6. files/6113/9171/8023/16_Abstract_ Mandel.pdf, accessed July 4, 2016

Join the discussion

Your email address will not be published. Required fields are marked *