More importantly, hyperalgesia to heat stimuli occurs as a consequence of surgical trauma even in the absence of RA, and is part of a well-recognized spectrum of post-incisional primary hyperalgesia that can last up to 7 days after surgery [17]. Secondary hyperalgesia is a similar response that occurs in the uninjured tissue surrounding the site of trauma. This represents the general phenomenon of peripheral sensitization to pain that is a normal physiologic response [18]. Tissue injury initiates a local inflammatory cascade, and the various inflammatory mediators (e.g., calcitonin gene-related peptide, cyclooxygenase [COX]-1, COX-2, prostaglandins [PGE], cytokines, interleukines, neurotrophins) activate peripheral nociceptors both at the site of injury and in surrounding tissues [19].

RA, and more specifically PNB, prevents the perception of pain by blocking impulse propagation in peripheral nerves from tissue nociceptors to second-order neurons in the dorsal horn of the spinal cord, and onward via ascending pathways in the lateral spinothalamic tract and subsequent thalamocortical pathways in the brain. As a result, RA will inhibit central sensitization to pain that is upregulation of the activity and responsiveness of spinal dorsal horn neurons [17,20]. However, PNB will not have a significant effect on peripheral sensitization, and this inflammatory process will continue unabated in the absence of systemically-administered medications [18]. Therefore, as peripheral neural blockade resolves, the nociceptive input from the hyperalgesic area at the site of injury will become apparent as rebound pain. This distinction between the effect of RA on peripheral and central sensitization may also be responsible for the lack of any observed association between acute rebound pain and the subsequent development of PPSP.

Laboratory research in cellular and animal models has reported several effects of local anesthetic administration that may affect acute nociception. Mice receiving sciatic nerve block with bupivacaine had microscopic evidence of early-phase peripheral nerve injury secondary to Wallerian degeneration and axonal demyelination [21]. Local anesthetics have also been implicated in neurotoxicity [22,23] and cytotoxicity [24] via disruption of mitochondrial membrane potentials and release of cytochrome C, accompanied by activation of caspases ultimately leading to cell apoptosis [24,25]. Proinflammatory effects such as COX-2 gene expression and subsequent increases in PGE2 production at the surgical site, as well as in cerebrospinal fluid, have been documented after local infiltration [26] and intrathecal [27] administration of bupivacaine, respectively. However, pain resulting from structural damage to neural tissue would be expected to be more prolonged than is typical of rebound pain. The relevance of these experimentally-derived neurotoxic and proinflammatory effects of local anesthetics to the clinical application of RA is therefore currently uncertain.

In summary, it appears unlikely that RA contributes to postoperative hyperalgesia to any clinically significant extent, and consequently it can be assumed that rebound pain does not represent an exaggerated nociceptive physiological response.

Cognitive and other higher-order cortical processes exert a powerful influence on the perception of acute pain. A cognitive bias [38] is a systematic pattern of deviation from rational judgment [39] that may lead to perceptual distortion. In the ‘contrast effect’ bias, a given stimulus is perceived as more intense when it is contrasted with a prior stimulus of lower intensity. The abrupt appearance of pain after a period of relative comfort, as the effect of a PNB wears off, can, therefore, cause a patient to rate the intensity of rebound pain higher than usual.

Patient expectations can also significantly influence pain perception [40–42]. Cumulative evidence shows that subjects who have been primed to expect good pain relief subsequently exhibit decreased pain perception and associated cerebral activity in response to noxious stimulation – a phenomenon known as placebo analgesia [43–47]. However, if the expectation of low pain intensity is not met, the disappointment may instead bias them towards reporting higher pain scores. This is relevant as patients who receive a PNB are often advised that they can expect excellent postoperative analgesia [13]; however, the finite duration of the sensory block may not be sufficiently emphasized and thus they are unpleasantly surprised by the pain that is unmasked.

Despite the issue of rebound pain, the use of RA for outpatient surgery results in increased patient satisfaction stemming from the avoidance of GA, effective postoperative analgesia with reduced opioid requirements, and decreased incidence of postoperative nausea and vomiting [49]. In a detailed study that interviewed patients who received PNB for ankle fracture surgery, Henningsen et al. [9] confirmed that despite the presence of rebound pain, patients reported high levels of satisfaction with RA and a preference for a similar technique in the future. These findings are mirrored in other studies that find that even though patients describe increased pain scores after PNB resolution, satisfaction scores remain high and similar to the GA-only group [2,8,50]. It, therefore, appears that from the patient’s perspective, rebound pain does not outweigh the early postoperative benefits of a pain-free interval [49], reduced opioid consumption and side-effects, superior recovery profile, and a shorter time to readiness for discharge [2].

Although poorly controlled acute postoperative pain has been implicated as a risk factor for the development of PPSP [51], there is no evidence to indicate that rebound pain per se predisposes to PPSP [13]. On the contrary, a recent Cochrane review reported that RA may instead reduce the incidence of PPSP after breast surgery and cesarean section [4]. As described above, the transitory nature of rebound pain, coupled with the early conduction block of nociceptive transmission, makes it unlikely that central sensitization will be exacerbated.

Rebound pain after RA has been implicated in higher rates of unanticipated health care resource utilization [13]. A retrospective study of 195 patients undergoing surgery for wrist fracture reported a higher incidence of unplanned physician visits (12% vs. 4%) because of severe pain in the first 48 h by those who received RA versus GA [13]. This may be partly explained by the fact that RA patients were far less likely to have received opioid and non-opioid analgesics prior to discharge, and there was no systematic patient education plan in place regarding post-discharge management of the postoperative transition from RA to systemic analgesia. A negative impact of RA (and the associated rebound pain) on healthcare utilization was not however borne out in a much larger retrospective study of over 59,000 patients undergoing outpatient shoulder surgery [52]. Patients who received a PNB, in fact, had a significantly lower rate of unplanned admissions, readmissions, or emergency department visits (9% vs. 12%) in the first seven postoperative days. Nevertheless, it is only logical that risk factors for rebound pain should be identified when performing RA for individual patients and strategies should be implemented to prevent and mitigate any potential impact on their postoperative recovery. This will be the focus of the remainder of this article.

The presence of preoperative pain [53] and younger age have been identified as patient risk factors for severe acute postoperative pain and PPSP [54]. Both of these have also been associated with a predisposition to rebound pain. Patients with pre-existing joint pain were more likely to report rebound pain following the use of PNB in total hip or knee arthroplasty [53]. Rebound pain following ankle fracture surgery with PNB as the primary anesthetic was also more common and more severe in patients under 60 years of age [55]. The mechanisms for this are incompletely understood, but include age-related differences in deep tissue (muscle) and superficial tissue (skin) nociception [56], increases in sensitivity of peripheral nerves to local anesthetics, and cationic acceptor site availability to local anesthetics [57], along with lower peripheral nerve conduction velocities in the elderly [58].

There is a perception that certain surgical procedures (e.g., shoulder, foot, and ankle) confer a higher risk of rebound pain. At present there is no data on the relative incidence of rebound pain among different surgeries; however, there are several logical contributing factors to consider. These include outpatient surgery, procedures associated with moderate-to-severe postoperative pain, and the use of single-injection PNBs (rather than continuous catheters) that provide near-complete analgesia [11,32]. Outpatient surgery patients are often discharged with a ‘one size fits all’ pain prescription that may be inadequate, and those who have received an effective single-injection PNB will usually have the first onset of breakthrough pain at home, where professional guidance and titration of analgesia is not available [13]. Patients often lack understanding of optimal medication dosing or timing and may also inappropriately limit the use of opioids at home because of fears of addiction or side-effects [13].

Rebound pain is a phenomenon that primarily manifests following PNB that provide dense sensory blockade (e.g., brachial plexus [2,7,10], popliteal sciatic [3,9] nerve blocks). Dramatic increases in pain scores and opioid consumption related to block offset are not usually seen in studies of fascial plane blocks such as transversus abdominis plane [59], pectoral nerves [60], erector spinae plane, [61] and quadratus lumborum [62,63] blocks. This requires further investigation for confirmation, but it may be related to factors that include a degree of visceral contribution to both pain and analgesic effect, an expectation of incomplete analgesic coverage by the block, and routine incorporation of these techniques into a multimodal analgesic regimen [64–66].

Extending the duration of sensory blockade to allow more time for healing and subsidence of the inflammatory process, as well as a less precipitous offset of block, should mitigate the impact of rebound pain. It is therefore not surprising that continuous catheter RA techniques with an infusion of dilute local anesthetic for 48 h or longer will preserve all of the early postoperative benefits of single-injection PNB while largely abolishing the phenomenon of rebound pain. Salviz et al. [67] randomized patients undergoing outpatient arthroscopic rotator cuff repair to receive GA alone, or GA combined with either a single-injection or continuous interscalene block. Compared to the GA-only group, both RA groups had shorter PACU stays, were discharged home earlier, and had a longer interval to first analgesic use. Most notably, the incidence of severe pain (8–10/10 on a numerical rating scale) on the first postoperative day was only 15% in the continuous interscalene block group, compared to 78% and 40% in the single-injection and GA-only groups respectively. By the second postoperative day, the single-injection and GA-only groups had similar pain profiles, but the continuous catheter group continued to exhibit lower pain scores with only 10% reporting severe pain compared to 35% in the other two groups. A similar effect was reported for continuous versus single-injection popliteal sciatic PNB in patients undergoing ankle fracture surgery. The peak in pain score trajectory was both delayed and attenuated in the continuous catheter group, and at 48 and 72 h postoperatively, pain scores were similar in both groups. However, the overall value of outpatient PNB catheters is controversial [68]. Continuous RA techniques are technically more challenging to perform, have an inherent failure rate [69], are time and labor-intensive to manage, and consequently are likely to remain under-utilized in this setting [68].

A more accessible alternative to continuous catheter techniques is the use of local anesthetic adjuncts to prolong the duration of single-injection PNBs. In a mouse model of sciatic nerve block with bupivacaine, the addition of perineural (but not intramuscular) dexamethasone prevented the appearance of a rebound hyperalgesic response to thermal stimulation [21]. Research indicates that perineural dexamethasone prevents bupivacaine-induced demyelination and Schwann cell degeneration [21], suggesting that any protective effect against rebound pain may be mediated by both anti-neurotoxic and anti-nociceptive mechanisms and effects. At present though, while it is well-established that dexamethasone (perineural more so than intravenous) [70] can prolong the analgesic benefit of PNB, there are no clinical studies specifically examining its impact on rebound pain per se.

Perineural buprenorphine is another local anesthetic adjunct used to prolong block duration, but again no studies have specifically investigated if it attenuates rebound pain compared to a control group. There is also a question of what constitutes an effective dose. In a retrospective cohort study describing their experience with a perineural combination of bupivacaine, clonidine, dexamethasone, and buprenorphine, Williams et al. [53] reported that a reduction in rebound pain after PNBs for total hip and knee arthroplasty was associated with > 300 μg buprenorphine but not lower doses.

Finally, although liposomal bupivacaine has been touted as an effective strategy to prolong the duration of analgesia (up to 72 h) with single-injection PNB [71], current evidence fails to support its routine use. Superior analgesia and opioid-sparing compared to conventional long-acting local anesthetics has not been demonstrated to date [72,73], and no studies have examined if it reduces the incidence and magnitude of rebound pain.

As discussed above, PNB only blocks the transmission of nociceptive input to the spinal cord and higher centers. Peripheral sensitization and other physiological responses mediated by the humoral inflammatory response to surgery remain unaffected. Combining RA with systemic multimodal analgesia is therefore recommended for the potential additive or even synergistic benefits [74–76] in improving postoperative pain and related outcomes. Nevertheless, many studies investigating rebound pain after PNB do not routinely incorporate perioperative systemic multimodal analgesia, and outpatient surgery patients usually receive significantly less analgesic medication prior to discharge [13,29,77].

Although there is no direct evidence that a consistent and comprehensive multimodal analgesic regimen will reduce rebound pain, it should be prescribed on a routine basis as part of good clinical practice [3,11–13,29,67,78]. This should include a combination of acetaminophen, non-steroidal anti-inflammatory drugs/COX-2 inhibitors, and oral opioids [28,53,67,79–81], in the absence of any patient or surgical contraindications.

As already mentioned, patients and caregivers should be clearly informed about both the advantages and limitations of RA. Day surgery patients should receive preoperative education on the finite duration of analgesia provided by PNBs, and depending on the surgical procedure, should specifically be warned to expect moderate/severe pain commensurate with the surgical procedure as the block wears off. They should be instructed to begin taking analgesic medication earlier rather than later, with an emphasis on the 15–20 min onset time for most oral analgesics versus the rapid offset of sensory block. A discussion of the expected interindividual variability [82] in block duration, pain thresholds, and response to analgesic therapy is also useful in assisting patients to self-titrate their medication. Supplementing verbal instructions with written or multimedia educational material will help improve compliance and lower perioperative anxiety and uncertainty [83].