THE practice of operative anesthesia is an application of Hippocrates’ dictum “primum non nocere ”—first of all, do no harm. A successful anesthetic is one that minimally interferes with the patient's homeostasis and does not add to the unavoidable quotient of surgical trauma. In this context, it is understandable that the notion of preemptive analgesia, by which early application of therapies may diminish later misery, is an enticing grail to pursue. The anesthesiologist's ministrations might then actually improve the patients state, rather than being at best a necessary evil. It is therefore a potential disappointment to conclude, as do Møiniche et al. 1in a detailed meta-analysis published in this issue of Anesthesiology, that there is little experimental support for a preemptive analgesic effect in clinical settings. They reviewed 80 randomized and blinded studies that compared various analgesic techniques applied before incision and later in the perioperative period. Only modest differences were noted, and these were present only with epidural injections.
The quest to identify a benefit from preinjury analgesic administration for surgical patients has been fueled by repeated and convincing demonstrations of this phenomenon in animals. The reason that a robust experimental finding cannot be confirmed in patients is not forthcoming, but several explanations can be considered. First, sensory blockade may not be adequate during surgery. The basis of preemptive analgesia is the prevention of increased responsiveness of central nociceptive pathways triggered by intense afferent neural activity. Even very brief sensory events can result in central sensitization, 2,3so effective prevention may require continuous sensory ablation throughout the surgical event. The intensity of afferent blockade is also important. Addition of systemic morphine to volatile anesthesia has no preemptive effect, 4whereas the more thorough action of intrathecal morphine does prevent central sensitization in rats. 5Small doses of intrathecal opioid show no preemptive effect, whereas larger doses have an amplified action given before injury, 6presumably through greater efficacy in blocking input from small nonmyelinated C fibers. 7The persistence of neuronal traffic, even during successful neuraxial anesthesia, 8may limit the preemptive effect of this modality, indicating the importance of thorough regional blockade. 9Increased strength of sensory stimulation may overcome the preemptive action of analgesics, even spinal local anesthetic, particularly if inflammation is a component of the injury. 10
Demonstrations of preemptive effects in experimentation on animals have used various types of injuries, including formalin injection and nerve trauma. However, in models that more closely emulate typical clinical surgery, the results are mixed. Spinal hyperexcitability 11and preemptive analgesia 12are evident in some studies of abdominal surgical injury, but there is no influence of analgesic timing on pain behaviors after peripheral surgery 13because of the minimal contribution of central facilitation. 14
The most obvious reason for diverging experimental and clinical findings is that animals may differ substantially from humans in pain pathophysiology and neuropharmacology. In rats, the species with which most studies of preemptive analgesia have been performed, sensitization is readily induced in spinal sensory pathways after conditioning stimuli. A large effect can then be seen when intense afferent activity is prevented from reaching the dorsal horn, which may not be the case in other species less prone to sensitization. Because genetic differences, even between various strains of rats, strongly affect the development of neuronal hyperexcitability 15and hyperalgesia 16after injury, comparison across species to human injury responses must be suspect.
It is not clear to me that the failure of preemptive analgesia is a great loss in the pragmatic clinical setting. Consider the outcomes used in studies of the topic. The analysis by Møiniche et al. 1tabulates pain scores, supplemental analgesic demand, and time to first postoperative analgesic. Pain scores, while important, should not differ if appropriate care is provided because postoperative pain in the absence of preemptive treatment is nonetheless responsive to adequate doses of analgesia, consistent with electrophysiologic observations in animals. 6A difference in supplemental analgesic demand should also not be an important medical issue. The need for 12 mg morphine in the recovery room instead of 4 mg before induction of anesthesia is itself not necessarily problematic. Although increased side effects from larger doses can be expected, there are no data to confirm this suspicion. Finally, a decreased time to first analgesic request in the absence of preemptive analgesia should not be a treatment problem, provided timely medication is available when the need arises.
The challenge, of course, is not to use the least amount of drug, but to minimize complications and optimize postoperative recovery. So far, there is minimal support for the belief that preemptive techniques aid recovery. One promising report confirms a favorable effect of early medication on the incidence and severity of chronic postoperative pain. 17In future studies of the timing of analgesic agents, it will be helpful to focus on aspects of recuperation, not only on initial postoperative pain levels or analgesic consumption.
Rather than being disappointed, I find the conclusions of Møiniche et al. 1to be encouraging. Trauma patients may be adequately treated with analgesics even though their injury has occurred without pretreatment because an early window of opportunity has not been missed. In scheduled surgery, the various side effects and complications that accompany intraoperative analgesic use may be avoided. For instance, intraoperative administration of opiates may lead to histamine release, dysfunction of the bowel, and disfunction of the biliary and urinary tracts, as well as acute opiate tolerance 18,19and hyperalgesia, 20which make postoperative pain treatment more difficult. Epidural local anesthetic block may complicate general anesthesia with hypotension and create diagnostic ambiguity afterward if paresis masks neurologic injury. For these reasons, initiating analgesic care at the time of emergence from general anesthesia may be more desirable; this is also a time when analgesic needs can be directly assessed. Because pain is “an unpleasant sensory and emotional experience,”21it cannot exist during general anesthesia. If preoperative and intraoperative analgesic treatment has little effect on the long-term course of sensory processes, their use before emergence lacks a clear motive.
There can be little doubt that great benefit will emerge from the burgeoning knowledge of processes underlying pain. Although experimentation may lead to drugs and techniques that can preemptively prevent pain in the clinical setting, our clinical attention must remain on treating pain when it presents, with adequate doses of proven agents.