“[Anesthesiologists are] in position to head off [postoperative] pain chronification, perhaps using our diverse pharmacologic armamentarium—if that is possible.”
A Classic definition of chronic pain is pain that outlasts the normal healing time.1 Implicit is the notion that pain mechanisms activated by injury should resolve along a predictable timeline and that chronic pain represents a failure of such resolution, i.e., “pain chronification.” Surgery is a prominent cause of chronic pain. Commonly cited estimates for the incidence of such pain suggest rates of between 5 and 85%, and chronic pain from surgery is a frequent reason for referral to pain clinics.2 Furthermore, anesthesiologists preside over the care of patients during the surgery and direct much of the pain-related care in the postoperative period. This seems to place us in position to head off pain chronification, perhaps using our diverse pharmacologic armamentarium—if that is possible. It is with great disappointment, then, that we read the comprehensive meta-analysis contributed by Gilron et al. that included 110 studies evaluating 28 different drugs representing 16 drug classes.3 The authors conclude, “Based on currently available evidence, none of the drugs studied so far can be recommended for clinical use specifically for the indication of preventing chronic pain after surgery.”
Arguably, the place to begin our examination of this failure is with an appreciation of the common causes of chronic postoperative pain. These are numerous and include musculoskeletal etiologies such as biomechanical disorders and contractures, neurologic etiologies such as neuromas, neuralgias and phantom pains, and an assortment of poorly understood syndromes such as complex regional pain syndrome, a well-known complication of surgery on the extremities. Neuropathic etiologies may deserve special attention since surgeries commonly associated with nerve damage, e.g., breast surgery, thoracotomy, and herniorraphy, all have high rates of chronic postoperative pain, and large outcome studies have linked neuropathic symptoms to more severe chronic pain.4 Likewise, the newer concept of “nociplastic pain,” pain attributable to altered nociception in the absence of ongoing tissue damage or lesion of the nervous system, may be important, as a clear etiology for chronic postoperative pain is often elusive.5 Nociplastic pain is characterized by pain sensitization. In fact, very few studies identified by Gilron et al. reported details of the pain phenotype, although such descriptions would likely demonstrate a highly variable pain experience. Considering the vast array of causes and features of chronic postoperative pain, however, it seems unlikely that individual preventative interventions would have the same degree of effectiveness against all of them. Failure to take these variables into account when designing trials and measuring effects may contribute to the equivocal outcomes.
Noting the fundamental problem of a lack of definition of chronic postsurgical pain, the International Association for the Study of Pain (Washington, D.C.) in conjunction with the World Health Organization (Geneva, Switzerland) developed a definition and classification scheme (box 1), and this new definition will be a part of the International Classification of Diseases, Eleventh Revision.6 This definition highlights a minimum timeframe of 3 months from surgery, distribution of pain plausibly relatable to the surgery, and absence of a separate primary cause, e.g., infection or loose hardware. This, then, is the preferred definition for chronic postsurgical pain and is the one to be used by clinicians and researchers moving forward. Until now, investigators have assessed chronic postoperative pain at timepoints from several weeks to 1 year postoperatively and have asked questions as vague as whether patients were experiencing any pain whatsoever and as specific as whether pain was experienced directly in the area of the incision. Problems resulting from the absence of a standard definition of chronic postsurgical pain contribute to the large variability in estimates of its incidence and the responsiveness to treatment. This situation is reminiscent of problems introduced by the absence of a standard definition of chronic postsurgical opioid use that is associated with a greater than 100-fold difference in estimates of incidence.7
Pain that develops or increases in intensity after a surgical procedure and persists beyond the healing process (3 months or greater).
This pain should
- Localize to the surgical field, or
- Project to the innervation territory of a nerve situated in this area, or
- Refer in a known pattern from this area
Preexisting pain conditions should be excluded
Other primary causes of pain such as infections, malignancy, nonunion of bone, loosening of surgical hardware, and other causes must be excluded
The new International Classification of Diseases, Eleventh Revision definition solves many problems related to the definition of pain but leaves in place others, particularly those surrounding the multidimensional nature of chronic pain. Ubiquitous visual and numerical pain rating scales measure pain intensity and have important roles in pain management and research, but they fail to address key outcome domains important for the diagnosis and management of chronic pain.8 The Initiative on Methods, Measurement, and Pain Assessment in Clinical Trials (IMMPACT) recommends several core outcome domains for use in chronic pain trials, including pain (intensity, qualities, and temporal aspects), physical functioning, emotional functioning, and, when studying treatments, patient ratings of improvement and satisfaction with treatment.9 It is intuitively understandable that having pain at a level of 4 out of 10 and functioning well is a less problematic situation than having the same pain score and not working, being disengaged from family, and becoming depressed. Including these additional outcomes may enhance our understanding of the value of our interventions.
There are several additional improvements in study design that could help advance the field. The first relates to the actual use of the pharmacologic therapies. Differences in drug doses, timing, and durations of administration hamper our interpretation of studies, as do issues related to the coadministration of nonstudy analgesics and interventions that might themselves impact rates of chronic postsurgical pain. For example, if an institution already uses a nerve block and ketamine infusion as components of a multimodal postoperative analgesic pathway, the effects of an additional agent may not be detectable. Moreover, it would be optimal to take into account established risk factors for chronic postoperative pain either during participant selection or as cofactors in analysis. One recently proposed framework suggests categorizing risk factors in six domains: genetic, demographic, psychosocial, pain, clinical, and surgical factors.10 It is clear that factors in some domains may be modifiable, e.g., psychologic status or surgical approach, but factors in other domains will not be, e.g., age and sex. Among these risk domains, pain status (preoperative and acute postoperative) is arguably the most strongly supported at this time. Another critical consideration is study power. Very few of the studies identified by Gilron et al. were powered for comparative measurements of chronic postoperative pain. This presents serious challenges as a study powered to detect a decrease from 20% to 10% of patients experiencing chronic postoperative pain with 80% power (P < 0.05) might require 400 participants. Finally, the measurement of harms caused by pharmacologic and other interventions, both major and minor, is a frequently omitted but important type of outcome that can profoundly affect our estimation of the utility of a treatment and acceptability of the treatment to our patients. These were poorly reported in most trials identified by Gilron et al.
One should view the conclusions of the Gilron et al. meta-analysis not as clearly demonstrating no preventative value, but rather as inconclusive despite a large number of patients being exposed to the experimental treatments. This is a problem largely related to heterogeneity and incomplete information in the component studies. Like many clinical questions critical to our field, much larger, more carefully designed trials will be necessary, and those are likely to require multi-institutional and even international cooperation. Attention to these factors might also enhance the generalizability of the findings. Not addressed in the meta-analysis were questions related to the utility of nonpharmacologic therapies that have been proposed to prevent chronic pain after surgery or combinations of therapies including various psychotherapies, neve blocks, prehabilitation/rehabilitation strategies, or neural stimulation. Regrettably, each of these fields is far less well-developed than pharmacologic prevention and likely suffers from the same pitfalls evident in the pharmacotherapy literature. Although we have a pathway to resolving important questions related to preventative strategies, for the moment, our use of pharmacotherapy to reduce the likelihood of developing chronic postoperative pain remains unproven.
Research support was provided by the National Institutes of Health/National Institute of Neurological Disorders and Stroke (Bethesda, Maryland; grant No. NS117340), the Department of Veterans Affairs (Washington, D.C.; grant No. RX001475), and the Department of Defense (Arlington, Virginia; grant no. CP190028).
Dr. Clark is a safety consultant for Teikoku Pharma USA, San Jose, California.