Anesthetics turn off the brain. At concentrations just slightly higher than anesthetic concentrations, the brain becomes isoelectric. Ponder that for a moment—it’s quite amazing! What other organ can we turn off and on in such a controlled manner? Surely not the heart, the liver, or the kidney. We can turn off, and then back on, not just any organ, but the human brain! Wow!

Our ability to turn off brain electrical activity invites the question: is this OK? The profound central nervous system effects of anesthesia raise concern that some untoward residual effect remains after the drug has been eliminated by the body. This concern recurs in the history of anesthesia, despite the apparent absence of any expression of concern from Gilbert Abbott, the first publicly anesthetized patient. Despite lingering concerns, over the next 150 years only minor unpleasant effects (eg, nausea and vomiting) persisted once the anesthetic left the body. The possibility that subtle impairment of intelligence might occur and persist was largely ignored until recently when the possibility of risk to young children was raised.

However, I was troubled by the question. My laboratory at UCSF had developed sensitive assays for many of the physiologic effects of anesthetics, but our focus on the effects on the mind had been limited to simple responses to stimulation (eg, response to voice or something painful).

Such simple responses seemed inadequate to judge the effects of anesthesia on cognition. Developing such tests would allow us to look for subtle levels of cognitive impairment following recovery from anesthesia. In the early 1970s, we collaborated with two psychologists who were experts in cognitive assessment.¹ We gave 40 volunteers 1.0 to 2.0 MAC halothane or isoflurane for durations ranging from 4.4 to 8.6 hours, each with or without 70% nitrous oxide. We looked for evidence of cognitive or psychological changes before and 2, 4, 6, 8, and 30 days after anesthesia. We used long exposures to anesthesia because we knew that the effects would likely be subtle. Were there a clinically important effect, we would mostly likely see it with a long exposure.

We subjected 41 unanesthetized controls to the same measurements and compared results for differences between the control subjects and subjects anesthetized with various anesthetics. We found that changes in function were greatest 2 days after anesthesia. Cognitive function had returned to near preanesthesia values 8 days after anesthesia. Thirty days after anesthesia we found only slight symptom and mood effects, and no intellectual effect attributable to anesthesia. Halothane produced greater negative effects on moods and symptoms, and tended to produce greater negative effects on intellectual function, than did isoflurane.¹

In sum, we found that prolonged and deeper levels of anesthesia can produce impaired cognition for up to 8 days after such anesthesia, and the choice of anesthetic may influence this effect. The study provided strong evidence that the effects of anesthesia on cognition were transient, and anesthetics did not produce lasting cognitive impairment.

The above studies in volunteers examined the effects of prolonged and profound anesthesia on psychological status and cognition, but they did not explore the effects of much smaller anesthetic concentrations. We knew nothing about trace exposure, the smallest concentration producing a measurable effect. Nor did they determine the residual concentration at which recovery is complete. I was worried about these, both because of the potential for harm to patients, but also because my colleagues and I spent every working day breathing trace concentrations of anesthetic drugs. Additionally, David Bruce and his colleagues had found evidence that trace concentrations²,³ were associated with impaired cognition.

To examine the effects of trace anesthetics and collaborate or refute the findings of Bruce and colleagues, Tom Cook and I used three tests of mental function (choice-reaction time; digit span test; Purdue Pegboard Assembly test) in human volunteers breathing alveolar enflurane or halothane concentrations less than 0.1 MAC⁴ or concentrations of halothane, nitrous oxide, or a combination of halothane and nitrous oxide less than 0.2 MAC.⁵ These low concentrations had no effect on any of our sensitive measures of mental function. These results confirmed similar results published by Smith and Shirley shortly before we published our findings.⁶ In my subsequent communications with David Bruce, he shared that his studies had experimental errors that he felt invalidated his results (D. Bruce, personal communication).

More than a decade later, about 1990, Hank Bennett called me. I didn’t know Hank, we had never met, and I was not aware of his interests. However, Hank had heard me present our evidence that there were no residual long-term anesthetic effects on mentation. Hank didn’t buy it. In his view, I had not used sufficiently sensitive measures of mental effects. Hank argued that I had ignored the possibility of long-term anesthetic effects on implicit (unconscious) memory. Sure, you can still drive a car, but can you still remember your violin lessons from third grade? Hank gradually convinced me that he might be correct. His arguments led to a productive collaboration that lasted a decade.

Rory Dwyer (Figure 13.1) led our first study, a study in which we measured MAC_awake. We provided our volunteers with answers to Trivial Pursuit-like questions (eg, what is the blood pressure of an octopus?).⁷,⁸ When we gave the answers to such titillating questions to awake patients, those answers were nearly always remembered. However, when we gave answers to patients breathing 0.6 MAC or greater concentrations of isoflurane, nobody recalled them when asked later. They stopped responding to simple requests, such as “open your eyes.” We also gave our modestly anesthetized patients suggestions for specific actions to undertake after they awoke. The anesthetized patients did not follow suggestions we gave them while anesthetized for things they should do when they awoke. Our team repeated the study with desflurane, with similar results.⁹

But there must be some subanesthetic concentration permitting the retention of information supplied. What might that be? We dared not study that in patients because it would not be acceptable to have patients awake during surgery. However, we could study that in volunteers!

And so we did. We found a 50% decrease in retention of Trivial Pursuit-like answers in subjects breathing 0.2 MAC isoflurane or 0.5 MAC nitrous oxide (Figure 13.2).⁷ Thus, concentrations of isoflurane approaching trace anesthesia significantly suppressed memory, just as Hank had suspected. We also found that isoflurane caused more amnesia than nitrous oxide in terms of relative potency measured by MAC (Figure 13.3).