How I started working in research

My first publication1 was the result of my interest in epidural continuous infusions. When I started working at Roswell Park Cancer Institute, I faced a significant number of patients with a history of opioid use who were experiencing, not only severe pain in the postoperative period, despite the utilization of a combinations of epidural bupivacaine and morphine at higher doses than opioid naïve patients, but also longer hospitalization time. Traditionally, opioid agonists like morphine, oxycodone, and fentanyl were believed to have equal analgesic effect when equivalent doses were used.2 However, Sosnowski and Yaksh showed non-symmetric tolerance response between intrathecal morphine and sufentanil in rats,3 introducing the concept of intrinsic efficacy, that is, the fraction of the receptor population each must occupy to produce a given effect.3 They propose that agents with higher intrinsic efficacy, such as sufentanil, will downregulate fewer opioid receptors over time when compared with agents will lower intrinsic efficacy, such as morphine. In parallel, there was knowledge that maximum analgesic effect may be achieved by these opioids while occupying different proportions of the Mu opioid receptor (MOR).4 Difference in receptor occupancy for the same effect are the basis for what we know now as intrinsic efficacy: opioids requiring low receptor occupancy are defined as having high efficacy. In fact, studies using non-competitive antagonists showed that sufentanil is superior to morphine when equimolar doses are compared.5 Thus, we theorized that in the face of MOR down-regulation, sufentanil would produce less of a right shift in the dose-response curve when compared with morphine. The corollary of that assumption is that patients with opioid tolerance would have a greater analgesic effect after sufentanil administration vs morphine use. In May of 1991, we faced a patient with a history of high preoperative opioid consumption (methadone 330 mg orally three times per day) who was crying due to severe pain in the postanesthesia care unit despite epidural bupivacaine/morphine therapy, we decided to test this hypothesis. The result was impressive; after administering 50 μg of sufentanil in 10 μg aliquots via the epidural catheter, the pain was controlled. It is noteworthy that we did not have knowledge of the value of intravenous ketamine in these circumstances or opioid induced hyperalgesia at that time. Nonetheless, the pain was controlled, but it returned in 2 hours. A second bolus of 50 μg of sufentanil was administered, followed by an infusion of sufentanil 2 μg/mL and 0.1% bupivacaine at 7 mL/hour. One hour later, she reported a pain score of 2/10 and remained below 4/10 throughout the postoperative period. On the fourth postoperative day, the patient consented to having CSF samples from the cisterna magna, the L4–5 intervertebral space, and blood. Sufentanil concentrations in the lumbar and cisterna magna were 0.34 and 0.19 ng/mL, respectively, and 0.28 ng/mL in blood. Without knowing it, we also documented the first case of the spinal and supraspinal analgesic effects of sufentanil at steady state. We reported the case6 and subsequently published a study comparing epidural bupivacaine/sufentanil therapy for postoperative pain control in patients tolerant to opioids and unresponsive to epidural bupivacaine/morphine.7 In this study, the results of the case report where corroborated. Twenty-one patients, age 31±6 years, undergoing major abdominal procedures for cancer were enrolled in the study during an 18-month period. Preoperative mean oral morphine equivalents usage was 380±97 mg/day (range 290–490) for 4±1 months. Before the switch to sufentanil, patients received 1800±432 μg/hour of epidural morphine for 5±2 hours and reported pain scores of 7–10/10. That amount is equal to 23±5 μg/hour of sufentanil. During the first 4 hours of sufentanil therapy, patients received 17±2 μg/hour and all subjects experienced pain control with pain scores <4/10. The difference in morphine versus sufentanil consumption was statistically significant (p<0.01), although it could be argued that the difference was not clinically significant. However, the difference in pain control, as judged by both the pain scores and the intravenous breakthrough opioid use, was significant. We concluded that sufentanil’s higher intrinsic efficacy appeared to have greater analgesic effects than morphine in patients with opioid tolerance.7

A year earlier, we also published a paper comparing the quality of analgesia and the hospitalization course of patients taking >50 mg of morphine (MS) per day in ppatients with cancer undergoing thoracic, abdominal, and hemipelvectomy surgery.8 It is noteworthy that currently, the Food and Drug Administration's (FDA’s) working definition of opioids tolerance classifies individuals with tolerance after the intake of 60 mg of MS per day (MME) for 1 week.9 The results showed that patients taking 183±99 mg of oral MS required 137±28 mg epidural morphine/0.1% bupivacaine and 48±4 mg of intravenous breakthrough MS for 218±42 hours compared with opioid naïve patients who used 44±15 mg of epidural morphine/0.1% bupivacaine and 10±6 mg of intravenous breakthrough MS for 76±35 hours for 76±35 hours.8 Thus, opioid tolerant patients, not only required larger doses of both epidural MS/0.1% bupivacaine doses and intravenous breakthrough MS but also a longer time to wean of the epidural therapy when compared with their opioid naïve counterparts. This observation helped us to justify a new treatment paradigm for patients with opioid tolerance that included increased MS epidural doses for a longer period.

Ready et al suggested that epidural MS was safe in the surgical wards when intermittent bolus of 4.2 mg and 3.8 mg every 8 hours were administered to maintain a stable level of analgesia in patients who underwent thoracic and abdominal surgeries.10 Since we did not have a 24-hour in-hospital service, we used continuous infusion for the same purpose since 1989 and our data base did not show a significant incidence of harms to our patients. Thus, we decided to publish our experience.11 Over a 4-year period, 4227 patients were treated with lumbar (53%) or thoracic epidural catheters (47%) for postoperative pain with solutions of 0.05%–0.1% bupivacaine plus 0.01% morphine at a rate 5 mL/hour titrated to maintain dynamic pain scores (pain during movement) <5%/10. 71% of the patients were treated on the surgical ward after the discharge from the postanesthesia care unit. Patients admitted to the intensive care unit (ICU), spent 1.2±0.8 days. Length of epidural therapy was 6.3±2.6 days. There were three cases of respiratory depression (0.07%), hypotension occurred in 3% of the patients, nausea/vomiting occurred in 22% of the patients, and opioid induced pruritus in 22% of the patients. We concluded that continuous epidural morphine/bupivacaine therapy is a safe and effective alternative to bolus injections of MS for the treatment of postoperative pain on the surgical wards.11 During that time, we also noticed that these patients were not only experiencing excellent analgesia, but also improved physiological outcomes and lower ICU complications. In a pilot study, we evaluated 203 patients undergoing upper abdominal surgery performed under light general/thoracic epidural bupivacaine/MS or general anesthesia and intravenous-patient controlled (IV-PCA) morphine analgesia, with two or more risk factors for coronary artery disease during a 2-year period, and determined the incidence of myocardial ischemia and myocardial infarctions.12 Demographic and clinical data were similar in both groups. Patients in the epidural anesthesia/analgesia group had a lower incidence of myocardial ischemic events (5 vs 15, p=0.04) and a lower number of myocardial infarctions but the sample size did not have the power to achieve statistical significance (0 vs 3, p=0.09). Likewise, we found that bowel recovery after radical hysterectomies performed with thoracic epidural bupivacaine/MS perioperative administration was faster when compared with those receiving general anesthesia and IV-PCA morphine analgesia.13 The epidural group had shorter nasograstric tube therapy, (4±3 vs 8±2 days, p=0001), tolerated solid foods sooner (6±2 vs 11±3 days, p<0.0001), and had a shorter hospitalization time (10±3 vs 14±4 days, p=0.0001) when compared with the IV-PCA group. We also documented these differences in critically ill too.14 In this study, 462 consecutive patients who underwent uncomplicated extensive surgeries of the thorax, abdomen or both with a duration of >3 hours were evaluated. Patients received epidural anesthesia and analgesia with combination of bupivacaine and MS via a continuous infusion or general anesthesia, followed by IV-PCA therapy with MS. There were no demographic differences between the two groups. Overall, 58% of the patients were admitted to the surgical ICU (SICU) immediately after the surgery. Patients in the epidural group required less days of ventilatory support (0.5±0.8 vs 1.2±0.9 days, p<0.05), less time in both the SICU (1.3±0.8 vs 2.8±0.6 days, p<0.05), and in the hospital (11±3 vs 17±5 days, p<0.05) than their counterparts. Moreover, when we calculated our per diem hospital reimbursement at that time, both SICU cost of care (US$1807.00 vs US$3892) and total hospitalization cost of care (US$4620 vs US$7140) was lower significant lower (p<0.05) for the patients treated in the epidural group.14 We theorized that a lower incidence of atelectasis, myocardial events, and duration of ileus were partly responsible for these differences, as the procedures were performed by only one surgeon for the thoracic procedures and by two surgeons in the abdominal procedures group. Consequently, we proposed that epidural anesthesia and analgesia techniques did not only provide better quality of analgesia, but also a lower incidence of complications that resulted in lower SICU and hospital stays.

Shortly after we published these results, studies with different outcomes were published. The MASTER trial,15 the Veterans Affairs Cooperative Study,16 and a study by Peyton et al 17 failed to show any benefits in the patients treated with epidural anesthesia and analgesia techniques. I was asked to write an editorial to the Peyton study18 and reviewed the other two studies as well. I argued that critical issues in the design and execution of these studies may be the reason why the studies failed to demonstrate a positive effect, and described the questions that I felt were important to determine if the use of the epidural anesthesia and analgesia was optimally implemented:

1. Did patients have a thoracic epidural placed according to the site of surgery, that is, T6–7 or T7–8 for upper abdominal procedures and T4–5 or T5–6 for the thoracic procedures?

2. Did they receive an intraoperative local anesthetic, and, if this occurred, was the administration continuous throughout the surgery?

3. Were there limits in the doses of inhaled anesthetics and intravenous opioids set for the intraoperative period to force the managing physicians into using an appropriate concentration and volume of an intraoperative local anesthetic?

4. What measures were taken to guarantee that the epidural catheter was indeed in the epidural space both during surgery and throughout the postoperative period?

5. Was a local anesthetic/opioid mixture used for postoperative analgesia, and was it used continuously throughout the study period?

6. How were uncontrolled pain and breakthrough pain defined and treated in the epidural group?

7. At what point during the day and how often was pain measured?

8. Did patients in the epidural group receive intravenous opioids, non-steroidal anti-inflammatory drugs, or both for the treatment of incidental/breakthrough pain? If so, how much?

9. Did the protocol determine criteria for extubation, and were patients extubated when they reached these criteria?

10. Did the protocol establish criteria for discharge from the ICU and the hospital?

As highlighted in the editorial,18 major concerns with the protocol design and execution in all these studies15–17 prevented their authors to demonstrate that perioperative epidural techniques decrease perioperative complications and improve outcome. A conclusion that I feel is still valid 20 years after its publication.

It is also noteworthy that the MASTER trail was underpowered for the primary outcome of death. In addition, the trial still found reductions in pulmonary complications.15

We were also intrigued by the results of several studies that showed that patients receiving epidural or intravenous fentanyl had comparable postoperative analgesia, incidence of side effects, fentanyl consumption, and plasma concentrations regardless of the route of administration.19–22 Moreover, it appeared that epidural lumbar or thoracic placement did not make a difference either.23 Consequently, we embarked on writing a review paper with the aim to explain the pharmacokinetic differences of frequently used opioids after epidural administration.24 We found that the octanol:buffer partition coefficients, and the meningeal permeability coefficients have a significant effect on the way opioids after epidural administration exert their analgesic action. Lipid solubility, as assessed by the octanol:buffer distribution coefficient, correlates with the meningeal permeability coefficient but in a non-linear fashion. The optimal octanol:buffer distribution coefficient that results in maximal meningeal permeability is between 129 (alfentanil) and 560 (bupivacaine).25 This biphasic relationship between lipophilicity and a drug’s meningeal permeability coefficient, may be explained by the dual nature of the arachnoid membrane which is the main barrier for drug diffusion between the epidural space and the intrathecal space.25 After a drug is deposited in the epidural space, but before it reaches the spinal cord, it must first cross a hydrophilic zone (extracellular and intracellular fluids) and then a hydrophobic zone (cell membrane lipids) of the arachnoid membrane.25 Thus, before there is diffusion through these two areas, the drug must first dissolve in those environments. Lipophilic drugs (ie, those with a high octanol:buffer partition coefficients) readily dissolve in the lipophilic component of arachnoid mater and thus cross the region easily. Conversely, they penetrate the hydrophilic zone with difficulty, creating the rate limiting factor in their diffusion through the arachnoid membrane. Drugs with intermediate lipophilicity move more readily between the lipid and the aqueous zones, and their meningeal permeability coefficients are correspondingly greater (eg, alfentanil, hydromorphone and meperidine).25 These physical and chemical properties of the opioids will also determine vascular permeability and fat sequestration in the epidural space, as it was shown by Bernards et al in a later elegant study.26

We also conducted a study to determine if postoperative epidural analgesia with fentanyl was related to plasma fentanyl concentrations in patients undergoing radical prostatectomies under balanced general anesthesia.27 On arrival to PACU, patients were randomized to receive a fentanyl bolus of 0.5–1.5 μg/kg, followed by a continuous infusion of 1 μg/kg/hour in either the epidural or intravenous routes via a PCA pump in a double-blind fashion. Bolus doses of 10 μg every 15 min with a 1 hourly maximum dose of 140 μg were provided. Infusions were stopped after 24 hours. Pain scores at rest and during activity were then assessed at 30 min intervals until rest pain scores were >7/10 (breakthrough time). There were no significant differences in total fentanyl dose utilization in the first 24 hours postoperatively (2479±139 vs 2583±207 μg, epidural vs intravenous, p=0.24). Likewise, pain scores at rest and during movement and plasma fentanyl levels between the two groups (mean 1.46 vs 1.28 μg, which are well above minimum effective analgesic concentration) were no different. However, median breakthrough pain, for both resting and dynamic VAPS, occurred earlier in the epidural group (120 vs 300 min, p=0.02). We concluded that the only explanation for these findings was the development of acute tolerance in the epidural fentanyl group. This conclusion was based on the research by Sofuoglu and collaborators,28 who showed that after inducing acute tolerance to opioids with morphine and administering norbinaltorphimine, a highly selective kappa opioid antagonist, 72 hours later, the ED50 of intraventricular morphine and other mu receptor agonists, was significantly greater in the mice that received the kappa antagonist than the control group. These results suggest that kappa opioid receptors play an important role in acute tolerance after mu receptor agonist administration at the supraspinal level. These findings are pertinent to our results because the antinociceptive action of epidural fentanyl is associated with binding of the opioid to both the kappa receptors at the spinal level and the mu receptors at the supraspinal level.29 30 Thus, based on Sofuoglu’s observations, we hypothesized that this interaction between the kappa and MOR at the spinal and supraspinal level may be one of the mechanisms underlying the acute tolerance experienced by patients receiving epidural fentanyl. More recent clinical studies have suggested opioid induced hyperalgesia after intraspinal or epidural fentanyl administration,31–33 and animal studies have proven it.34–38 Nitric oxide generated from the inducible nitric oxide synthase has also been implicated.34 Since the diagnosis of acute opioid tolerance vs opioid induced hyperalgesia is done after the fact, it is also possible that what our patients experienced was the later—that is, opioid induced hyperalgesia.

The interest in cellular mechanisms of opioid tolerance and how that knowledge translated to the management of patients with opioid tolerance in the perioperative period evolved dramatically in 10 years, since our initial publications on this subject. Thus, we reviewed the subject and highlighted the following practice points.39

1. Opioid tolerance may occur as early as 2 weeks after therapy is started with opioids. Again, the FDA’s working definition of opioids tolerance classifies individuals with tolerance after the intake of 60 mg of MS per day (MME) for 1 week.9

2. Patient treated with high doses of opioids preoperatively may have a greater analgesic effect when treated with another opioid with higher intrinsic efficacy, such as sufentanil.

3. Basic research supports the role of the NMDA receptor in the development of tolerance. Thus, the use of an NMDA receptor antagonist, such as ketamine, may play an important role in the perioperative pain management of these patients.

We continued to manage our patients with history of opioid use with these practice points in mind.40

In parallel with the work in opioid tolerance in the perioperative period, we were also interested in cancer pain management. That led to a publication on the use of neurolytic superior hypogastric plexus block for chronic pelvic pain due to cancer.41 We found that the block was effective in providing good pain relief, while decreasing opioid consumption in 69% of the patients. However, we learnt that in those patients where the cancer had extended beyond the viscera, that is, evidence of lymphadenopathy, or growth in the peri pelvic areas, the blocks were not successful. At that time, we thought that this could be due to an incomplete neurolysis of the plexus, because all these patients failed to show adequate spread when contrast medium was injected. However, after having the same experience with splanchnic nerve neurolysis in upper abdominal cancer painpatients, we now believe that incomplete pain relief in these patients who have both somatic and neuropathic pain components, cannot be treated by neurolytic blocks of the sympathetic axis.42 The afferent fibers innervating the viscera in the abdomen travel with the sympathetic nerves, trunks, ganglia, and rami.43 Thus, by performing a sympathetic neurolytic block, the fibers conducting somatic and neuropathic pain signals are not affected. We subsequently replicated these findings in a multicenter study and since then have avoided neurolytic blocks of the sympathetic axis when there is evidence of tumor spread beyond the visceral capsule.44

In medicine, research is performed to shed light on unanswered questions. However, rarely, we focus on unquestioned answers. Studies have shown that the incidence of postdural puncture headaches is lower after a dural puncture with a cone-shaped tip needle such as the Greene, Whitacre, and Sprotte needles when compared with a Quincke-Babcock needle. The rational is that the former ‘spread’ the fibers of the dura mater, whereas the later cut the dura sharply.45 However, this was disproven by work that we published back in 2000.46 We evaluated dural lesions produced by 25-gage Quicke and Whitacre needles with scanning electron microscope and found that Whitacre needle produced explosive lesions both in the epidural and subarachnoid side of the dura mater, creating flaps of collagen fiber at the edge of the lesions with a ripple effect around the hole. In contrast, the Quincke needle produced clear cuts in the dura mater, regardless of the direction of the bevel in relation to the axis of the spine.46 Considering these findings, we suggested that the inflammatory reaction created by the lesion after the use of a Whitacre needle was responsible for creating a ‘plug’ that resulted in a lower incidence of postdural puncture headaches. Regardless, based on these findings, we recommended that perpetuating the concept of ‘spreading’ fibers with the cone-shaped tip needles is not accurate because scanning electron microscopy of the human dura mater showed that its fibers neither run in longitudinal direction, nor are arranged in a parallel fashion.47

In the same line of thought, prior to the advent of MRI compatible leads and generators, there were safety concerns when performing an MRI in patients with a conventional spinal cord stimulator (SCS). These concerns were based on the static magnetic field, the static magnetic field spatial gradient, the gradient magnetic field, the radiofrequency field generating rotational forces (torque) and the so-called missile effect on the device with potential tearing of surrounding tissues. Moreover, current induction by the electromagnetic field may result in device malfunction or failure, and the radiofrequency-induced currents in device heating and patient burns. Consequently, we sought to determine the safety of performing an MRI in patients with implanted SCS at different levels in the spine in a prospective in vivo study.48 MRIs in different areas of the body were performed with a 1.5-Tesla clinical use magnet and a specific absorption rate of no more than 0.9 W/kg. We found that, under the conditions of this protocol, MRIs in patients with SCS systems resulted in few complications, which were not serious, and in no case were patients harmed or the systems reprogrammed. All patients reported satisfaction after the studies.

Even though the results of SCS in the treatment of various cancer pain conditions did not have a long-term analgesic effect, peripheral nerve stimulation has been very successful in providing long-standing analgesia in patients with postsurgical nerve injuries.49 In this article, the applications were presented: Auriculotemporal and lesser occipital nerve for persistent post craniotomy pain, V1 and V2 trigeminal subdivision for persistent postenucleation of the eye pain, superficial cervical plexus for persistent postmodified radical neck pain, intercostal for persistent postthoracotomy pain, intercosto-brachial for persistent post-mastectomy pain, ilioinguinal, genitofemoral, and iliohypogastric for persistent postlymphadenectomy at the groin area were described. Since then, we have also implemented this therapy for persistent postamputation toe pain. There were also contributions in practice parameters for the use of SCS in the treatment of chronic pain,50 neuropathic pain,50 intrathecal therapy,51 the management of cancer related pain,52 chronic pain management,53 and acute postoperative pain.54 The acute pain guidelines work also generated a paper on the research gaps found on the drafting of the Practice Guidelines for Acute Postoperative Pain Management.55 Moreover, the impact of perioperative pain management on cancer recurrence was also addressed.56

Being faithful to the American Society of Regional Anesthesia and Pain Medicine Mission of Education

On 2017, JAMA published a paper—the MINT study—on the effect of radiofrequency denervation (RFD) on patients with chronic low back pain.57 Their objective was to evaluate the effectiveness of RFD added to a standardized exercise program for patients with chronic low back pain with a positive diagnostic block at the facet joints, sacroiliac joint, or a combination of facet joints, sacroiliac joints, or intervertebral disks, who had failed conservative care when compared with the standardized exercise protocol (control group). They found that RFD combined with the exercise program did not improve the pain in the population studied when compared with the control group. JAMA published three letters to the editor addressing some of the concerns with the design and execution of the study. Vorobeychik et al 58 addressed the reference to the Spine Intervention Society guidelines by Juch et al,57 in support of the protocol used to select patients for RFD. They highlighted that they used less that 50% pain reduction after a single block of the posterior ramus medial branch blocks, which has been associated with a 25%–45% false-positive rate.58 Moreover, they expressed the same concerns with the criteria used for the diagnostic block of the sacroiliac joint. Rimmalapudi et al 59 also addressed the concern with the less than rigorous single diagnostic facet medial branch block and the use of mean pain intensity rating scale measurement. They asserted that this method may produce biased results that relate to the distribution of pain reduction in population studied and worse, because of the use of 30% pain improvement to define treatment success. This threshold could have enhanced the success rate in the exercise group. They also criticized the use of 22 g needles for the RFD, as they result in almost 50% smaller in the volume of the lesions when compared with 18 g needles.59 Kao et al 60 had interesting comments about the use of minimal clinically important difference (MCID) of at least 2.0 points on an 11-point Numeric Rating Scale for low back pain, which had been proposed by Ostelo et al 61 and highlighted that this patient-reported outcome should not be used to compare groups that are analyzed in a cross-sectionally approach because they were created to define a threshold for change within individuals in a longitudinal form. Ostelo et al warned against the use in comparing group changes.61 Interestingly, Ostelo was one of the authors of the MINT study.61 Kao et al also noted that the FDA warns in their Guidance for Industry60 that ‘group average is not an appropriate threshold for individual change’. Thus, the appropriate analysis should have been ‘the determination of change within each individual group longitudinally followed by a comparison of the mean numeric pain scores at baseline vs 3 months in the two groups or, alternatively, comparing the proportion of patients who experienced more than a 2.0 points pain reduction between groups’.60 Considering these important issues, The American Society of Anesthesiology and Pain Medicine (ASRA) Board Members felt that, in line with our mission, we should highlight the study design and performance flaws of this study to contribute to the education of our members.62 Moreover, we were concerned that national expenditures for spine conditions in the USA have increased on average 7% per year63 without a proportionate reduction in the prevalence of backpain, spine surgery, or disability rates.63 64 Consequently, we reasoned that education to perform these procedures in a stringent manner was critical. Aside from the issues already addressed in the letters to the editor, we highlighted the flaws in using a perpendicular approach to the medial branch when using a conventional thermal RFA lesion, as a lesion develops horizontally along the shaft of the needle, with very little tissue lesion will occur at the tip of the needle.62 Moreover, we expanded on the criticisms by Rimmalapudi et al 59 regarding the use of a 22-gage needle by providing a schematic diagram illustrating the size of a monopolar RF lesion on a lumbar medial branch with a parallel to the nerve placement. For an 18 g cannula with a 10 mm tip, the mean effective lesion radius is 1.6±0.5 mm and the margin of error, under these circumstances would be 0.6 mm. Thus, recognizing both the smaller size of the lesion and the margin of error when using a 22 g cannula, limiting the heating period to 90 s, and using a single lesion produced with a perpendicular approach is critical to understand the flawed results of the study.62 In other words, the likelihood that patients enrolled in the study for treatment with this RF protocol had an optimal mean surface area lesion were very low. If one adds to this technical problem the validity of the use of (MCID) for group comparisons, and patient selection, it is not surprising the lack of difference between the two groups. Unfortunately, our intentions were misconstrued by Lanier and Neal.65 We addressed their misguided position by pointing out that our intention was to promote education and analytical thoughts as to why we felt that the study was flawed, to establish the basis for an open discussion on these flaws, and as a corollary of these two positions, to maintain access for patients with facet related and sacroiliac joint pain. We were very concerned that, as a direct result of this publication, reimbursement for these procedures ceased to exist in the Netherlands.65 To this end, we pointed out that the American Society of Regional Anesthesia and Pain Medicine had a policy to respond to scientific quality concerns in publications dealing with regional anesthesia, acute pain, and chronic pain. Moreover, since our ethical standards were called into question, we also reminded them that the Cochrane guidelines for systematic reviews welcome the input from practitioners who are experienced in the examined practice and that we were alarmed that they ignored the technical procedural questions that we addressed because they lack the professional experience to do so.66

Research to hopefully help my colleagues

The burn-out syndrome was first described by Freudenberger.67 The syndrome arises when other strategies fail to overcome the occupational stress.68 The individual affected may experience serious physical and psychological consequences that are directly related to the intensity and duration of this status if the individual cannot develop adaptive mechanisms to restore the lost psychological equilibrium. Moreover, behaviors that may be deemed unprofessional are one of the strategies that the individuals affected by this problem use to cope (see depersonalization below).68 In the proposed definition by Maslach and Jackson in 1986 and subsequently revisited in 2001,69 the syndrome consists of three dimensions: emotional exhaustion, depersonalization or more clearly, dehumanization, and low personal accomplishment at work. The emotional exhaustion dimension is self-explanatory, the realization that there is no energy left to carry out daily work activities.70 Depersonalization manifests with emotional detachment from people to whom the professional should be caring for and this behavior may also extend to coworkers. The interactions become impersonal, insensitive, and uncaring. Individuals may also show demeaning, harsh, cynical or ironic behaviors when dealing with patients. This dimension is considered the defensive element of the syndrome.70 The last dimension is personal achievement, and it usually reflects a decrease in loss of satisfaction and efficiency at work that frequently becomes a burden. The Maslach Burnout Inventory (MBI)69 has been recognized for more than a decade as the best instrument to measure the probability of burn-out based on the extensive research done with it since its initial publication over 25 years ago.69

The incidence of burn-out among professionals experiencing high levels of stress has been reported to be around 10%.71 72 However, there is a large landmark study that evaluated the incidence of burn-out in physicians in the USA. Questionnaires were sent to physicians and 7000 responses were obtained (27% of the requests sent for evaluation). These responses were then compared with a probability-based sample of 3442 working US adults. Physicians were more likely to have symptoms of burn-out (40% vs 28%) and to be dissatisfied with work-life balance (40% vs 23%). The results were highly statistically significant with a p<0.001 for both measurements.73 Anesthesiologists scored just above the mean for rates of burn-out.73 There is no consensus as to what the best way is to treat this problem. However, there seems to be agreement in that prevention is the key and recognition of its existence is vital. Previous studies have documented that the most affected domain among anesthesiologist is personal achievement.71 72 An incidence of low personal achievement ranging from 36% to 48% has been documented in these studies, which correlated reduced staffing in the face of increasing workload and external pressures to increase productivity as the main culprits for the low scores in this area. Low personal achievement has been hypothesized to lead to depersonalization (dehumanization) which in turn reflects a high probability of burn-out. Consequently, the first step is recognition of a feeling of low personal achievement to avoid depersonalization and burn-out. You may ask, why the need of awareness if we need to continue to work and there is a low likelihood that working conditions will change. Burn-out has been correlated to musculoskeletal disorders and cerebrovascular disease.74 Moreover, there is an association between burn-out and sick-leave absences, disability, and admissions for mental health issues, back pain, and lack of restful sleep.75 76 Thus, our well-being may be in jeopardy if we do not pay close attention to this potential problem because of the increases in workload, longer working hours, and demanding patients with aggressive, belligerent behaviors.

Considering these issues, a study to determine the incidence of burn-out among anesthesiology subspecialty society members was started. Study participants entered voluntarily via website access provided by the ASRA, the Society for Pediatric Anesthesia, and the Society of Cardiovascular Anesthesiologists. We asked members of the subspecialty society practicing acute and chronic pain management, pediatric anesthesiology, and cardiac anesthesiology to take the MBI-Human Services Survey (MBI-HSS), the Veterans RAND 12-item Health Survey, and the Social Support and Personal Coping Survey. Multivariable regression analysis compared the groups, and adjusted burn-out prevalence was compared with an all physician and an employed general population sample. Among 1303 participants (response rates 21.6%–35.6% among the subspecialty groups), 43.4% met established burn-out criteria (range 30.0%–62.3%). Chronic pain physicians had significantly worse scores than the other three groups of subspecialty anesthesiologists, the all-physician comparator group, and the general population comparator group. Mental health inversely correlated with emotional exhaustion and depersonalization in all groups. Self-identified burn-out correlated with the full MBI-HSS (R=0.54; positive predictive value of 0.939 (0.917, 0.955)). Physicians’ scores for personal accomplishment were higher than population norms.77 The chronic pain practitioners had worse emotional exhaustion (low physical and emotional energy levels) and depersonalization (loss of empathy that manifests as cynicism) scores than the regional/acute pain, pediatric, and cardiac subspecialty groups (p<0.001). In fact, 62% of the chronic pain practitioners had burn-out identified by the MBI-HSS scores, and 38% of those were aware that they were experiencing burn-out. Thus, it was disconcerting that 24% were not aware that they were experiencing this condition.77 The Medscape National Physician Burnout and Suicide report published in 2021 reported that 51% of critical care practitioners were experiencing burn-out and that constituted the highest rate among the 29 specialties.78 Anesthesiologists were reported to have a 40% incidence,78 which is a number like the one reported in our study for the other subspecialties, except cardiac anesthesiologists (56%).77 The report also pointed out that 26% of these individuals cope with burn out by drinking alcohol,78 which adds another layer of concern. Equally important, despite the increased awareness, the proportion of pain physicians who met established criteria for burn-out is high and is more frequent than comparable US non-physician employees.79 The findings in our study have significant implications as it has been shown that physicians experiencing burn-out deliver care that is less safe and less patient centered.80 Consequently, as pain physicians face more pressure from decreased reimbursement for professional services, government agencies focus on erroneous patient satisfaction intermediate outcomes, such as Press-Gainey surveys that do not reflect the quality of care that we deliver,81 and hospitals continue with a misguided attempt to improve healthcare focus on making people happy rather than healing them,82 we will likely see a higher risk of professional burn-out. Consequently, the focus of a pain physician will need to include, not only strategies to remain up to date with the new developments in the literature, but also his/her well-being.