Gadgeteering for Pain Relief: The 2021 John W. Severinghaus Lecture on Translational Science

This structure of the lecture was based on lessons from Joseph Campbell and The Power of Myth, a 1988 Public Broadcasting Service documentary series in which Bill Moyers interviewed Joseph Campbell, an author and scholar of myths across time and cultures.²  Filmed at George Lucas’s Skywalker Ranch, where Campbell had helped Lucas in creating the first Star Wars film, it’s a fascinating conversation and well worth seeing. Campbell kept returning to what he termed the hero’s journey, a series of events that are remarkably similar in the life stories of the Buddha or Shirley Chisholm or Ulysses or Jean Valjean or Jesus, and in books and films for children and adults today.

One of the most common hero’s journeys can be described largely in three phases (fig. 1A). The first starts with leaving the comfort of the familiar to achieve a goal. Next, the hero overcomes, aided by strangers, mentors, and the experience of many trials, and is transformed through these events—in some cases without even achieving the goal. Finally, the hero returns, bringing new life to society through the hero’s own vitality and experience of life’s mystery. You and I are not mythical heroes, but we can follow similar steps in aspects of our lives, including our working careers.

As an example, consider the plenary lectures at the 2008 ASA Annual Meeting, which appeared as articles in the April 2009 issue of Anesthesiology and were highlighted on the cover. Ronald Miller, M.D., went to Vietnam, returning with leadership and new knowledge of resuscitation and pharmacology to advance the specialty. Steven Shafer, M.D., went out to mathematical modeling, returning with concepts shaping drug development across medicine. And John Severinghaus, M.D., went on to become a physicist, returning a physician with devices that transformed medical care. Herein, I describe my own journey in parallel with that of Severinghaus and invite you, the reader, to consider your own personal journeys as we progress together.

We are a remarkably young specialty. Dr. Ralph Waters was founding our specialty when John Severinghaus was a boy. This is how John remembered Dr. Waters: “He [Dr. Waters] lived about two blocks from us in Madison. I played with his kids and his dad. His office was next to my dad’s office at the university. Between the two, there was a drinking fountain, a bubbler, as we called it in Madison. And one of my very early memories was of Ralph Waters taking a drink with his pipe still in his mouth from the bubbler between the two offices.”³  Physicians from around the globe came to Madison, Wisconsin, to study under Ralph Waters and returned to establish, as chairs, new departments of anesthesiology, birthing the specialty.

Figure 2 depicts some of the characters at the start of the story. In his 30s, John Severinghaus was recruited into anesthesiology by Robert Dripps, M.D., at the University of Pennsylvania (Philadelphia, Pennsylvania) and was subsequently recruited to the University of California, San Francisco, on the condition that an independent Department of Anesthesiology be created, headed by Stuart Cullen, M.D. As a resident in my 30s, I planned on entering private practice until the morning I heard Tony Yaksh, Ph.D., present an early morning grand rounds lecture at the Mayo Clinic (Rochester, Minnesota). He’s still my mentor. My first chair, Francis M. James III, M.D., remains a mentor, sponsor, and role model to me to this day. This article shows images of many old white men mentioned in this talk, some of them young who became old. This reflects our specialty’s history and my life experience in a position of privilege in our caste system. I should have done better.

Act I is all about achieving a goal. It begins on a clear day, and the goal, obviously requiring hard work to reach, is visible and the paths to it clear. For many early-stage scientists and certainly for me, the goal translates to “I will show” (fig. 1B). The focus is to prove a concept or create something to address a problem. Gadgeteering is most evident in Dr. Severinghaus’s career during this phase. In the early 1950s, he invented an esophageal probe for physiologic monitoring, which he considered to be an unfortunate commercial flop.

The rationale for other gadgets during this period came from urgent needs presented by the polio epidemic. As he states in the Wood Library-Museum of Anesthesiology interview with Thomas Hornbein,³  “I had built an electro-phrenic respirator the last year of medical school after watching Stan Sarnoff demonstrate how he could stimulate his wife’s phrenic nerve just by touching the surface of her neck with this probe. I thought that would be a great way to do artificial respiration. This was really before we had much in the way of ventilators: the iron lung. It was a polio epidemic and that looked like a great bet for doing artificial ventilation. […] The three electrodes and the three techniques in blood gases, all in a sense came from this common need to measure blood, oxygen, CO₂ and mostly in the face of the polio epidemic.” After decades of maiming and killing children and young adults, the polio epidemic ended with near universal vaccination, a true miracle of modern medicine. Rapid blood gas analysis went on to fundamentally advance perioperative and critical care medicine.

While researching this lecture, I’ve come across many stories about Severinghaus the gadgeteer. My favorite, contributed by the Swedish anesthesiologist Torsten Gordh, Jr., was Severinghaus’s construction of a suspended pipe and an electric toy train to ferry arterial samples from a third floor operating theater to the laboratory, also on the third floor, in an adjacent building in Leeds, United Kingdom, for Professor John Nunns.

In my own Act I, I planned to show that α2-adrenergic agonists could treat labor pain safer than existing methods. Lumbar epidural analgesia using local anesthetics was the primary technique during my medical training, but case reports of untreatable cardiac arrest⁴  and paralysis⁵  appeared, leading to new Food and Drug Administration (Silver Spring, Maryland) regulations and drug withdrawals. Also, case reports of respiratory arrest and death from neuraxial opioids began to appear.⁶  I submitted a grant proposal to the National Institutes of Health (Bethesda, Maryland) as a resident, arguing that α2-adrenergic agonists might be safer drugs to administer for obstetric anesthesia. That grant (No. R23GM35523 [Epidural Clonidine Analgesia in Obstetrics: Sheep Studies]) was funded during my obstetric anesthesia fellowship, and continued for the next 27 yr.

I was initiated into basic science in perinatology in the laboratory of James Rose, Ph.D., Professor of Obstetrics at Bowman Gray School of Medicine (now Wake Forest School of Medicine [Winston-Salem, North Carolina]). The commonly used animal model for obstetric physiology research at the time was pregnant sheep: probes and catheters were inserted to assess uterine blood flow and maternal and fetal hemodynamics and arterial blood gases. We used a Radiometer (Copenhagen, Denmark) arterial blood gas machine that was only a couple of generations beyond the one John Severinghaus first created, one of which currently resides in the Smithsonian Institution (Washington, D.C.). We described drug transfer from the ewe to the fetus and disposition of clonidine in plasma and cerebrospinal fluid (CSF) after epidural, spinal, and IV administration. This led to interactions with Donald Stanski, M.D., Chair of Anesthesiology at Stanford University (Stanford, California), and Steven Shafer, a friend and colleague ever since. Clonidine produced minimal or no hypotension in pregnant and nonpregnant sheep, so I showed what I had hoped to show.

At this time, I also received a grant from the orphan products division of the Food and Drug Administration in which I planned to show efficacy of epidural clonidine in humans. I would have failed if not for the intervention of two strangers, a classic event in the hero’s journey. After obtaining investigational new drug approval, we began a randomized, controlled trial of one dose of clonidine for analgesia after surgery. My first study patient had new-onset atrial fibrillation, which resolved spontaneously in the intensive care unit. I called the orthopedic surgeon, Gary Poehling, M.D., to say we were stopping the study. He felt I had no reason to believe this was study-related and encouraged me to continue the study, which I did after Food and Drug Administration review. After the first three subjects in the study had no analgesia, I was again ready to stop. While stewing over this, I traveled to Hamburg, Germany, to present some of our sheep research, and Tony Yaksh invited me to discuss science over a beer at a nearby pub. We were joined by a world-renowned pharmacologist from Memorial Sloan Kettering Cancer Center (New York, New York), Charles Inturrisi, Ph.D. He asked whether any of the patients had side effects from the dose of clonidine I had chosen, and I said no. This brought a hearty laugh and a 30-min explanation of the concept of dose-ranging studies. I returned home, redesigned the study, and showed dose-dependent analgesia from clonidine.⁷ 

Unfortunately, clonidine turned out to be a complete flop for obstetric anesthesia. Spinal clonidine does produce pain relief for 90 to 120 min in women during labor, as does epidural clonidine after cesarean delivery, but analgesia is accompanied by unacceptable hypotension and sedation.^8,9  Clonidine was an equal flop in nonpregnant patients after surgery.¹⁰  In contrast, in patients with chronic regional pain syndrome, infusion of lumbar or cervical epidural clonidine for lower or upper extremity disease reduced pain compared to placebo from severe to moderate levels and was well tolerated.¹¹  In a randomized, controlled clinical trial in patients with intractable cancer pain, epidural infusion of clonidine produced sustained analgesia that was sustained for weeks,¹²  leading to Food and Drug Administration approval. However, the Food and Drug Administration appropriately added a black box warning stating that “[E]pidural clonidine is not recommended for obstetrical, postpartum, or perioperative pain management.” I clearly did not show what I wanted, but this work and that from others did expand the use of α2-adrenergic agonists from nasal decongestants and antihypertensive pills to a new, nonopioid treatment for cancer pain, a useful sedative, and an adjunct to regional anesthesia.

Severinghaus had a mind that was open to new ideas: “I began to look for easier ways to measure CO₂. I went to the physiologic society meeting in Madison in 1954 and heard Richard Stowe describe a carbon dioxide electrode and it clicked just like that, and I said, that’s what I need.”³  During Act I, I too heard just what I needed at meetings, including two international meetings I organized about α2-agonists. Lawrence Saidman, M.D., invited me to join the Associate Editorial Board of Anesthesiology, and others must have recommended me to serve on National Institutes of Health grant review committees and an Food and Drug Administration Advisory Board and to give plenary lectures at international venues. I suspect Dr. James opened doors at the American Society of Anesthesiologists, the Association of University Anesthesiologists (San Francisco, California), and the American Board of Anesthesiology (Raleigh, North Carolina) and probably others. Our research group was small, with mostly European physicians, including Astrid Chiari, M.D., who gave birth to her first child when we were at the World Congress in Vienna, Austria. She later became Chair of the Department of Anesthesiology at the University of Vienna. Barry Seltzer, M.D., a musician and internist, and I established a book group of two every Tuesday night that has lasted to this day as we help each other navigate different stages of career and life. Like John Severinghaus, I took a sabbatical during this time, with 6 months in Iowa City, Iowa, where I became known as Sweet Baker James as I woke up early to bake sweets for the graduate students and fellows in the laboratory, then 6 months in Paris, France—a truly fantastic time for both career and family.

Act I involved what appeared to be mostly serendipity, like the two strangers who helped me in that first clonidine study. Just 2 weeks ago, I serendipitously came upon that orthopedic surgeon as I was bicycling home. I hadn’t seen him in more than 20 yr, and took the opportunity to tell him the role he played in my life. Act I also involved many “invisible hands,” a term Bill Moyers mentioned, to which Campbell replied, “When you begin to deal with the people who are in the field of your bliss they open doors to you.” I will return to the concept of bliss later. Taking time such as a sabbatical is critical in Act I to truly experience and reflect on this bliss that is central to your journey.

Act II is when the path becomes murky, dangers occur, and one can get lost or discouraged. It’s moving away from being confined inside a rigid box filled with unconscious biases in “I will show…” toward a freer space of, as Steve Shafer likes to say, “letting the data speak.” In both of our journeys (Severinghaus’s and my own) during this phase, we examined spinal fluid; he was seeking an understanding for adaptation to high altitude, and I was manipulating spinal cord processing on pain transmission.

To do this, we gadgeteer-ed on ourselves as well as others. John Severinghaus’s early work involved spinal fluid sampling, as well as respiratory measurements and arterial and jugular venous sampling, at the University of California Barcroft High Altitude Laboratory (Bishop, California). He described his postdural puncture headaches and his approach to prophylactic treatment: “I think I’ve had a headache with every needle tap except two. One of them was done at night with me lying face down. And I never turned over so that the needle hole was the highest thing in me from that time on all night. And I didn’t get a headache after that one, but I tried that a few more times. Didn’t work! And finally, when I was doing another one of these, must have been in the mid-1970s, I had 1 cc of blood taken from a vein and injected into the needle into the CSF as the needle was being pulled out to leave some fibrin inside. Didn’t get any headache after that one!”³  In my own case, I opted for the more conventional epidural injection of blood to patch the hole and relieve the headache.

As a byproduct of the work during this period, we studied spinal drug distribution, in part to better understand the pharmacology of novel spinal agents and in part to address a perplexing clinical observation with spinal anesthesia. With similar patients and injections, distribution of spinal anesthesia is rapid and extensive in most, but slow and limited in some. This has resulted in manipulations of injection method, volume and baricity, patient positioning, and maneuvers like coughing or straining to alter distribution. As a resident, I noticed that each attending had a special recipe they swore by, arguing, like confident chefs, that all others were wrong. This didn’t seem very satisfactory.

We solicited the help of Steve Shafer and included pharmacokinetic/dynamic modeling in several studies of drugs injected neuraxially. For opioids, we inserted a needle in a low lumbar interspace and another more cephalad, injected fentanyl through the caudad needle, and sampled through the cephalad needle as drug spread from the injection site during the next 2 h with both needles in place (fig. 3). The average CSF fentanyl concentration from the cephalad needle (black line) peaked by 15 min and then slowly declined (fig. 3; black line). One individual (fig. 3; red line) had near-maximal concentrations at 1 min after injection. This very rapid movement is a good thing with local anesthetics, but could underlie cases of acute respiratory depression 10 to 20 min after spinal injection of lipophilic opioids during labor.¹³  Another individual (fig. 3; green line) had no fentanyl in CSF from the cephalad needle until 4 min after injection, and the fentanyl concentration in CSF at 10 min was 40-fold lower than the one shown in red. This could predict the maddeningly slow distribution of spinal anesthesia with local anesthetics in some cases, making all of us anxious—especially the patient.

Steve Shafer successfully modeled these curves with a mixing kinetic constant.¹⁴  Every individual had a different constant and we could not predict it from subject size, sex, age, or lumbosacral CSF volume measured by magnetic resonance imaging. We wondered if something else might predict it. CSF circulates in the cranium from the third ventricle around the cortical surface, where it is absorbed by arachnoid villi. Textbooks describe a similar lazy river–like flow of CSF in the spinal space, but there is little evidence for such bulk flow. Rather, studies dating back more than 50 yr suggest that CSF pressure fluctuates with each cardiac cycle¹⁵  and that CSF jostles up and down, like water under a toilet plunger, as it is displaced from the cranium when blood volume increases then decreases. We thought that someone with a larger cardiac stroke volume might have faster and more extensive drug mixing in CSF than another person with a smaller stroke volume, and showed that the duration of spinal fentanyl analgesia during labor increased as stroke volume, assessed by pulse pressure around the time of injection, also increased.¹⁶  This hypothesis deserves further testing.

We also explored drug distribution and action of novel drugs in spinal analgesia. For clonidine, we determined the transfer rate from the epidural space to CSF, generated computer-controlled epidural infusions to target constant CSF concentrations, and calculated CSF concentrations associated with analgesia.¹⁰  For neostigmine, we injected a drug in a lumbar interspace, leaving the needle in place for 2 h for CSF sampling; Steve Shafer successfully modeled the spinal cord analgesic effects in the foot and in the hand from CSF levels at the injection site, using a concept he called observation from a distance.¹⁷  The only problem was that every individual had a very different CSF concentration curve, making it useless for clinical prediction.

Funding during this period was based on a National Institutes of Health grant (No. R37GM48085) that is still active after nearly 30 yr, with Alison Cole, Ph.D., as my program director. Alison retired in 2020, and the ASA and Foundation for Anesthesia Education and Research thanked her in person for her service to many anesthesiologist–scientists at presentations of the 2021 ASA Annual Meeting. The goals of these studies were to test whether drugs producing analgesia in animals also worked in humans. These studies paralleled or followed those of Swedish physicians including a Ralph Waters trainee and longtime friend of John Severinghaus, Torsten Gordh, Sr., M.D., who introduced anesthesiology to Sweden and lidocaine to the world¹⁸ ; Narindar Rawal, M.D., who helped introduce morphine¹⁹ ; Torsten Gordh, Jr., M.D., who introduced clonidine²⁰ ; and Alf Sollevi, M.D., who introduced adenosine²¹  for neuraxial use.

The first drug we tested was built on anatomic and physiologic studies showing that neurons in the locus coeruleus project to the spinal cord, releasing norepinephrine, which causes analgesia by stimulating α2-adrenoceptors (fig. 4A). Systemic morphine activates this circuit, and it’s now believed that norepinephrine reduces pain directly by an α2-adrenergic mechanism and indirectly by stimulating spinal acetylcholine release, which is also analgesic (fig. 4B). In a clinical study for another purpose, we amended the protocol to test this idea: I stayed in the research unit overnight with a spinal catheter in place and received IV morphine on the next day. Norepinephrine increased 2.5-fold and acetylcholine increased 10-fold in my CSF after IV morphine, peaking at the time of maximum effect.²² 

These and other studies led us to “wonder if…” spinal neostigmine would be analgesic in humans via acetylcholine-augmenting effects and if systemic opioids and epidural clonidine would be potentiated by this mechanism. To test neostigmine’s effect on IV opioid analgesia, we gave alfentanil by IV computer-controlled infusion to targeted plasma concentrations and showed that the concentration response curve for analgesia was shifted up and to the left, meaning it was potentiated, when spinal neostigmine was administered.²³  To test neostigmine’s effect on epidural clonidine analgesia, we gave clonidine by computer-controlled epidural infusion to targeted CSF concentrations and showed that neostigmine also potentiated clonidine.²⁴  Several dozen clinical trials of neuraxial neostigmine have been performed over the past 25 yr. Intrathecal injection produced analgesia in these trials, but also multiple side effects, especially severe nausea and vomiting, and its use was abandoned. Epidural injection, on the other hand, does not cause nausea and vomiting, but its use is limited to that of an adjunct to local anesthetics.

During the next 25 yr, we asked “I wonder if…” with spinal injection of other compounds, examining efficacy, side effects, pharmacokinetic/dynamic models, and mechanisms. A strong noxious stimulus releases enough glutamate into the spinal cord to stimulate N-methyl-d-aspartate (NMDA) receptors, which signal pain and drive sensitizing processes (fig. 4C). For this reason, we examined spinal amitriptyline, which, in addition to its monoamine reuptake properties, is an NMDA receptor antagonist. Incoming pain signals also stimulate cyclooxygenase-containing cells in the spinal cord, which release products that drive spinal sensitization (fig. 4D). For this we reason, we examined spinal ketorolac to block spinal cyclooxygenase. Some descending pain inhibitory systems release enkephalins and other opioids, which directly stimulate opioid receptors and indirectly stimulate adenosine release for analgesia (fig. 4E). For this reason, we examined spinal adenosine. Finally, painful input results not only in activation of descending systems to the spinal cord that block pain, but also some that augment pain and sensitization. The clearest example from animal work is serotonin (fig. 4F). For this reason, we examined spinal ondansetron to block serotonin receptors and stop pain amplification.

Is this kind of work gadgeteering or creation? Severinghaus had this to say: “An aha! is a sudden insight into something that you hadn’t seen connect before. […] The nature of genius perhaps, or of creation, is when our mind suddenly connects two things that had no connection before in a way that’s creative and new and makes something work that didn’t work before. Now, I didn’t have an ‘aha!’, I just took their devices and worked with them and built them.”³  Perhaps he’s too hard on himself, but I, too, felt that my work during this time might be useful, but was not particularly creative.

Table 1 shows the steps we used for testing: replication (Could we replicate an analgesic effect in animals in our own laboratory?); toxicology (Is the drug toxic to the spinal cord?); efficacy; side effects in the analgesic range; and clinical utility as the balance of efficacy and side effects. For clonidine, there is utility in the restricted population of intractable cancer pain. For neostigmine, there is no utility spinally (shown in red), but there is some epidurally (in green). Notice that only two of the other four made it to human studies, and neither of those showed clinical utility: adenosine, which had low efficacy in only one type of pain; and ketorolac, which had no efficacy in any type of pain. Was this worth 30 yr of work? Spinal clonidine and perhaps adenosine have a place in treating intractable cancer or chronic pain, and our studies with neostigmine and adenosine did spur development of cholinergic and purinergic receptor agonists and modulators, so, maybe yes. The two drugs that did not make it to humans also advanced our understanding.

First, in a multicenter study with Tony Yaksh, we failed to replicate analgesia from spinal ondansetron in rats.²⁵  There was a growing realization at this time that many findings in basic research in prestigious journals could not be replicated, and that this reflected very low scientific quality. John Ioannidis stated that these research findings may often be simply accurate measures of the prevailing bias,²⁶  and I wrote an editorial showing this to be the case with studies of low scientific rigor I myself had published early in my career.²⁷  An article that applied principles of evolution to this concern noted that “…an incentive structure that rewards publication quantity…will lead to the natural selection of poor methods and increasingly high false discovery rates.”²⁸  The reasons for this are many and are entrenched in basic science culture. At Anesthesiology, we responded with a policy to ensure that authors at least report basic aspects of study design, such as blinding and randomization, but we have far to go to fix this problem.²⁹ 

Second, we showed neurotoxicity from spinal amitriptyline in rats that had been missed in previous behavioral studies.³⁰  Tony Yaksh has studied safety of spinally administered drugs for his entire career and advised the Food and Drug Administration on spinal drug development. He often quotes a founder of pharmacology, Paracelsus, who stated in 1532, “All things are poisons. It is only the dose which makes a thing poisonous.” Dose was the reason women were paralyzed by accidental spinal injection of 2-chloroprocaine intended to be epidural.⁵  Tony also showed that concentration and time of exposure explained paralysis in patients when lidocaine was slowly infused through a spinal microcatheter.³¹  He showed that all NMDA antagonists are spinally toxic in dogs³⁰ ; others observed similar toxicity from spinal ketamine in people.³²  At Anesthesiology, we adopted a policy not to publish clinical studies of neuraxial or perineural injection of drugs not Food and Drug Administration–approved for those routes unless there was Food and Drug Administration oversight, or there was toxicity screening in the literature, or there was a long history of use (see the journal’s online Instructions for Authors). We applied this policy as an example of the central ethical concept in medicine to first do no harm. This policy has now been adopted by many journals in our specialty and beyond.

Toward the end of Act II, I, like Severinghaus, was asking broader questions that required entire teams. The National Institutes of Health supports team science with large grants, and after 2 yr of effort, we secured funding for a team of chronic pain clinicians, physiologists, receptor biologists, and neuroscientists, supported by an external advisory board of national pain research leaders to study what we called “Pharmacologic Plasticity in the Presence of Pain” (National Institutes of Health award No. P01-NS41386). We asked why some drugs like opioids lose efficacy with chronic treatment whereas others, like clonidine, show better efficacy for chronic than acute pain.

My own laboratory was at its largest at this time, with fellows who would achieve leading research positions in their home countries of France, Japan, Korea, the United Kingdom, and Hungary, or who would immigrate to the United States to leading positions. I told new fellows in the laboratory at the time that it was a teaching laboratory, that education was its primary purpose, and that science happened as a byproduct. Looking back on this time, I was kidding myself in that I was learning from them as much as I was teaching. As Joseph Campbell states, “Just as the people who you met by chance became effective agents in the structuring of your life, so you have been an agent in the structuring of other lives.”²  This was also true in my work at Anesthesiology, where I taught and learned from people at different stages of journeys within and outside our specialty. Perhaps the greatest honor I have received in my work life was to give the Rovenstine lecture in 2015,³³  and the audience and I learned so much more by including Anesthesiology Associate Editor Carol Cassella in the creation and performance of that lecture.

This period was very unsettling to me, and perhaps to Severinghaus, as he thought about his electrodes. “I’m afraid I have to admit that that’s what everybody remembers me for. And it probably will be, if anything, it goes down in history as being my creation. That doesn’t make me particularly happy that that is it, because I would have liked to think that some of my physiologic ideas have been more important. I’m having more trouble selling that notion.”³  During this time, I felt that the goal became less clear, the path more obscure, and the meaning less certain. Some of our clearest findings, such as toxicity (which had been missed by others), lack of replication, and lack of efficacy in humans, seemed to have no impact on the entrenched culture that primarily valued new targets, new circuits, and new mechanisms. I also wondered whether my work might actually cause harm, like previous thought leaders who encouraged deadly practices like epidural injection of 150 mg bupivacaine over 2 min for cesarean delivery, or 10 mg epidural morphine for postoperative analgesia with only 30 min of respiratory monitoring. I have since come to terms with this risk of harm, as all scientists must, but nearly left research altogether during this period. Two mentors, colleagues, and friends, Douglas Ririe, M.D., Ph.D., and Timothy Houle, Ph.D., listened, encouraged, and supported me in finding my path forward through Act II.

This brings us to Act III, the return, in which the physician–scientist asks, “I wonder why…?” As Severinghaus said, it’s about hypotheses: “I think that a paper which not only does research and reports it, but creates a hypothesis which is testable and stimulates other people is perhaps the most important thing you can do. And that’s what I really would like to be remembered for.”³  Clearly he is remembered for this. Dr. Beverley Orser, the 2020 Severinghaus Lecturer, mentioned to me that she witnessed firsthand the tremendous gift John had to ask himself and others the right questions to generate those kinds of studies.

When asked about the future of arterial blood gas analysis, John Severinghaus said, “There is a fiber optic catheter that can be slipped in through a 20-gauge needle into an artery and give you pO₂, pCO₂, and blood pressure. Now, having said that, I don’t think it’s going to be a big success because it’s invasive.”³  I moved away from invasive spinal injections for the same reason.

The question I’m now asking is about recovery from pain after surgery. We need to start thinking about this in a different way. When our daughter, Laurel, was in labor with her first child, I texted Dr. Pamela Flood, who had created new mathematical models for labor progression. She sent me a link, where I entered Laurel’s age, weight, race, and gestation, and plotted each cervical dilation as reported to us by her husband. Dr. Flood’s model precisely predicted Laurel’s labor course. Obstetricians would scoff at not considering labor a process. Yet we have only rudimentary knowledge of the time course of recovery from pain after surgery—we just say pain is chronic or gone. To better understand this, we measured daily pain for 2 months in several hundred women after cesarean delivery; our modeling showed five patterns.³⁴  When Laurel later had a cesarean delivery, she sent me pain scores for a few days, and I accurately predicted the group she belonged to and her time to pain resolution.

There are other pain syndromes where we do consider recovery as a process. A few years ago my father self-diagnosed a painful rash in a low cervical dermatome as shingles, even though he had received the shingles vaccine. I reassured him that, like the current vaccine for COVID-19, the first-generation shingles vaccine he had received was documented to reduce the likelihood and severity of acute disease and speed recovery, and it did for him. In looking for a vaccine against chronic postsurgical pain, we need clues for the tremendous interindividual variability in patterns of recovery from pain. This is the “I wonder why…?” question.

One clue is the admonition, which appears to be well founded, that you have to work through the pain to get better. This conflicts with modern pain theory that repeated acute pain, such as each step after a knee replacement, causes pain chronicity rather than recovery. To sort this out, we spent a few years gathering preliminary data and brought together a multidisciplinary team of basic and clinical researchers and an external advisory board and received a large team science grant in 2016 that ends in 2022 (National Institutes of Health award No. P01-GM113852).

The underlying hypothesis is simple (fig. 5A). Each step after knee surgery signals pain to the spinal cord and is transmitted to higher centers. Some higher centers send signals back to the spinal cord. These don’t dampen each jolt of pain, but over time, they do help resolve the pain-amplifying sensitization in the spinal cord caused by surgery. In rats, postoperative sensitization is measured by a low threshold to withdrawal from touch, which recovers in about 2 to 3 months, similar to recovery time from postoperative pain in humans. If the descending norepinephrine fibers are destroyed, recovery is blocked.³⁵  We’ve replicated this finding several times, suggesting that this system is critical to recovery in rodents. We then developed a noninvasive way to quantify the transient activation of the norepinephrine system in people via acute pain stimulus and are testing two hypotheses in a multicenter clinical trial with Daniel Sessler, M.D.

The first is that a larger norepinephrine response to acute pain, as measured by our noninvasive test, will predict quicker recovery from pain after knee or hip arthroplasty during the course of 6 months. The second relates to the clear failure of gabapentin to speed postoperative recovery from pain. In 2021, a meta-analysis examined 110 studies in more than 19,000 patients, concluding that the effects of gabapentin were small and of uncertain clinical relevance.³⁶  We have repeatedly shown that gabapentin increases tonic firing in the locus coeruleus in rodents,³⁷  and this makes us think that we might be able to predict patients for whom gabapentin would produce a benefit. In cardiac physiology, we know from the Frank–Starling relationship that IV fluid will improve or worsen stroke volume in patients with heart failure, depending on the starting end-diastolic volume (fig. 5B). Others have shown a similar relationship between tonic activity in the locus coeruleus and its acute response to a stimulus. We hypothesize that gabapentin, by increasing tonic locus coeruleus activity, will increase the acute response with each step during recovery in some individuals, and decrease it in others, depending on the level of tonic activity before gabapentin (fig. 5C). Based on this, we are testing whether knowing each patient’s tonic activity in this system before surgery will allow us to predict preoperatively who might benefit from gabapentin with quicker recovery and who might be harmed by gabapentin.

The other clue comes from the observation that there is a faster recovery from pain after childbirth, including cesarean delivery, than after other major surgical procedures.³⁸  More rapid recovery from injury also occurs in rodents and can be reversed by blockers of oxytocin,³⁹  which is normally released into the blood and into the spinal cord during labor and for weeks after delivery. After several years gathering preliminary data, we formed a multidisciplinary team of investigators from Stanford to Sweden under the leadership of T. Jeff Martin, Ph.D., with advice from experts from Germany to San Francisco to test how oxytocin speeds recovery from injury in rats and to prepare for clinical trials.

Oxytocin can act peripherally and centrally to relieve pain, but we’re really interested in the periphery because of a new story. Normal input from peripheral nerves, like when you’re pinched, results in pain. But abnormal input, like light brushing on sunburned skinned, can also cause pain. It’s assumed that peripheral nerve fibers are injured during surgery, providing abnormal input and pain, and that this resolves in a few weeks, so any pain after this must reflect abnormal central sensitization alone. In our laboratory, Mario Boada, a wizard at neural recording and the smartest physiologist I know, has shown instead that surgical injury affects both pain and touch fibers and that the time-course of their recovery exactly parallels behavioral recovery during the months⁴⁰  after surgery, meaning it’s mostly abnormal input, not central changes. More excitingly, he’s shown that oxytocin can reverse the abnormal peripheral input causing prolonged pain.⁴¹ 

We know little of peripheral oxytocin pharmacodynamics outside of obstetrics, and little about pharmacokinetics due to problems with historical assays. Again, we relied on Steve Shafer’s help to move us forward, determining oxytocin pharmacokinetics in blood with a better assay, developing noninvasive measures of peripheral and central oxytocin action in men and women, and creating preliminary kinetic/dynamic models after computer-controlled IV oxytocin infusions to fixed plasma concentrations. These models predict that an infusion at a fixed plasma concentration over minutes can produce primarily peripheral effects, that a brief infusion at a higher concentration can produce primarily central effects 12 h later, and that a sustained infusion at a much lower concentration over days can produce a primarily peripheral effect. We will test these predictions over the next 5 yr to allow proper design of clinical trials after surgery. We will also study intranasal oxytocin, a more practical route for prolonged treatment.

You cannot imagine the joy of this part of my journey, returning vitalized and changed, sparking creativity together with other wonderfully vital scientists. Mythical journeys are usually walked once, whereas each of us starts afresh on paths each day or year. In science, these journeys include the assertion “I will show…” and the questions “I wonder if…?” and “I wonder why…?” It’s worth noting that another door opened for me at the start of this phase, and it has been my honor to serve with a remarkable team, the Board of Directors of Foundation for Anesthesia Education and Research and my joy as we go about supporting investigators as the curtain rises on their plays.

In preparing this talk, I came across a photo of me in the laboratory office and noticed a book sitting by my shelves. It’s one of Piet Hein’s small books of haiku-like sayings or poems, which he called gruk. Hein was a Danish physicist, architect, and artist, and perhaps because he was all these things at once, John Severinghaus was drawn to him and would often quote a gruk in his scientific articles. At the end of announcing his good friend John Severinghaus as the recipient of the ASA Excellence in Research Award in 1986,⁴²  Thomas Hornbein ended with what he considered to be John’s favorite gruk⁴³ :

I believe John did know what it was all about because he experienced it himself. Joseph Campbell, a Buddhist at heart, questioned whether there was any purpose or meaning in life, but was convinced that myths gave us clues to being truly alive as one is following their bliss. To Campbell, following one’s bliss meant participating fully in life in what he called “a wonderful opera, but painful.”²  To me it’s evident that John Severinghaus was following and experiencing his bliss. As I ended the lecture with a photo of John Severinghaus and his wry smile, I could almost hear him repeating the words of Campbell: “If you do follow your bliss you put yourself on a kind of track that has been there all the while waiting for you, and the life that you ought to be living is the one you’re living. I say follow your bliss and don’t be afraid, and doors will open where you didn’t know they were going to be.”²  Like each of us, John Severinghaus had many doors opened for him, just as he opened many for others. He would—as I do—want each of us to go out and open doors for others.

The author dedicates this lecture to his father, John Eisenach, M.D., a contemporary of John Severinghaus who attended the lecture virtually from his home, and who has inspired the author as a World War II veteran, a family doctor and then anesthesiologist, a pilot, a singer in retirement, and an exemplary and loving father to this day. The journey of a physician–scientist’s career reflects many helping hands, visible and invisible. The author notes that, if along the way he has listened to better understand others, if he has been kind or generous along this journey, it’s because of the love and example of his mother, Minnie, and wife, Patricia.

The author thanks Judith Robbins and Maureen Robst at the Wood Library-Museum of Anesthesiology (Schaumburg, Illinois) for providing many images and materials for this lecture. The work described in this lecture was performed under grants GM35523 (R23, R25, then R01; 1985 to 2008), GM48085 (R01, then R37; 1992 to 2022), P01-NS41386 (2002 to 2007), and P01-GM133852 (2016 to 2022) from the National Institutes of Health (Bethesda, Maryland).

Supported in part by currently active grants (Nos. R37 GM48085 and P01-GM113852) from the National Institutes of Health (Bethesda, Maryland).

Dr. Eisenach is President of the Foundation for Anesthesia Education and Research, an organization related to the American Society of Anesthesiologists (Schaumburg, Illinois). Dr. Eisenach holds U.S. patents 60/356,280 (“Compositions and Methods for Treating Pain Using Cyclooxygenase-1 Inhibitors”) and 6,248,744 (“Method for the Treatment of Female-specific Pain”) related to the content of this article. Products have not been, nor are currently, in development under these patents.