A systematic review of therapeutic interventions to reduce acute and chronic post‐surgical pain after amputation, thoracotomy or mastectomy

1Chronic pain may occur in a minority of patients after minor operations, but the incidence is much greater after procedures that involve major nerve injury or transection. This systematic review examines the impact of specific interventions on the management of perioperative pain and associated chronic pain in high-risk operations such as amputation (50–80%), mastectomy (20–40%) and thoracotomy (30–50%; Kehlet et al.,; Macrae, ). Neuropathic pain is typically due to a damaged or dysfunctional nervous system (Woolf and Salter,; Macintyre and Schug,; Gray, ) and it may be characterized by sharp, shooting or burning sensations associated with sensory loss and paradoxical hypersensitivity (Macrae, ). The underlying pathophysiological processes are highly complex, but involve the interaction of the immune and nervous systems, which results in peripheral and central sensitization (Scholz and Woolf,; Harvey and Dickenson, ). Neuropathic changes are traditionally associated with chronic pain, whereas acute pain has been considered as a protective nociceptive process. However, acute neuropathic pain is now also considered to be an eminently treatable but significantly under-recognized condition (Dworkin et al.,; Gray, ). The true incidence of acute neuropathic pain after surgery is uncertain (Gray, ), but the incidence of acute neuropathic phantom limb pain after amputation is greater than 50% (Jaeger and Maier,; Nikolajsen et al., ). In addition, severe post-operative pain is associated with a far higher risk of developing chronic pain (Kalkmann et al.,; Bisgaard et al.,; Kehlet et al.,; Hanley et al., ). Therefore, prompt, aggressive management has the potential to make a substantial difference in the short term and, perhaps even more importantly in the long term as well before symptoms become entrenched. However, if perioperative pain management regimes only incorporate opioids, non-steroidal anti-inflammatory drugs (NSAIDs) and regional analgesia, then patients with acute neuropathic pain may be undertreated and therefore at significantly increased risk of developing chronic pain. Chronic pain is considered to be a complex biopsychosocial condition and a long-term multidisciplinary approach to management may be required for many cases. The therapeutic options available for established chronic pain have substantial limitations and problematic side effects; therefore, preventative techniques are of particular importance.

2Medline (1946–January 2014), Embase (1974–January 2014) and the Cochrane Library.Thirty-two randomized controlled trials met the inclusion criteria.

3Medline (1946–January 2014), Embase (1974–January 2014) and the Cochrane Library.

4Thirty-two randomized controlled trials met the inclusion criteria.

5Appropriate dose regimes of gabapentinoids, antidepressants, local anaesthetics and regional anaesthesia may potentially reduce pain.Ketamine and intercostal cryoanalgesia were not effective at reducing chronic pain.Total intravenous anaesthesia may be beneficial.

6Appropriate dose regimes of gabapentinoids, antidepressants, local anaesthetics and regional anaesthesia may potentially reduce pain.

7Ketamine and intercostal cryoanalgesia were not effective at reducing chronic pain.

8Total intravenous anaesthesia may be beneficial.

9A literature search was carried out using Medline (1946–January 2014), Embase (1974–January 2014) and the Cochrane Library. The research question was, ‘What is the evidence for the use of specific therapeutic interventions in the management of perioperative pain in procedures that have a high risk of chronic pain and whether these interventions reduce chronic post-surgical pain?’ We used a combination of keywords and subject headings. The subjects of interest included amputation, mastectomy, thoracotomy, perioperative, pain, neuropathic, chronic pain, pre-emptive, epidural, regional and the relevant drug names and classes. All subject headings were examined for relevant terms that are related to the types of operation performed and the types of intervention. Articles were limited to randomized controlled trials (RCTs) in adult human participants. Additional studies were identified by manually tracking references from published papers. On some occasions, where a manuscript was unavailable, the authors were contacted directly.

10Studies were also excluded in which subjects were recruited after the perioperative period and those with follow-up periods of less than 2 months were excluded. Studies with high risk of bias as assessed by the Cochrane Risk of Bias Tool (Higgins and Green, ) were excluded. Additional studies were identified by manually tracking references from published papers. On some occasions, where a manuscript was unavailable, the authors were contacted directly.

11All of the trials were reviewed independently by each of the authors and consensus was reached as to the inclusion and interpretation of each individual trial. This review focused only on surgical procedures associated with a high incidence of chronic pain: amputation, mastectomy and thoracotomy. Studies investigating acute pain and chronic pain were analysed, whereas studies in which subjects were recruited after the perioperative week and those with follow-up periods of less than 2 months were excluded. Once all of the trials were collated and analysed, an assessment was made of the potential to perform a meta-analysis of the quantitative efficacy of specific interventions.

12The combined literature search identified 598 records and a further 6 additional records were identified through the manual tracking of references. After the duplicates were removed, 331 records were screened, and of these, 252 excluded as being clearly irrelevant. A total of 79 records were therefore assessed for eligibility and of these 47 were excluded (the most common reasons were as follows: not RCT, inappropriate study design, high risk of bias, inadequate size, not related to pain and/or chronic pain not assessed). Thirty-two trials met the eligibility criteria and were included in the qualitative analysis.

13Of the 32 trials, 5 (16%) studied amputation, 17 (53%) studied thoracotomy and 10 (31%) studied mastectomy. Details of all studies are summarized in Tables 1 and 2. The mean sample size was 89 (±12 SD) subjects (range 36–343). The mean duration of patient follow-up in the trials was 7 (±1 SD) months following the surgical procedure (range 2–15). Four of the trials had more than two study arms and (reflecting the complexity of clinical practice) 13 trials incorporated multiple interventions within the same study. The outcome measures varied between studies; the visual analogue scale (VAS; 0–100 mm) was used in 22 trials; a version of the numerical rating scale (NRS; 0–10) was used in 20 trials; and the verbal rating score (0–3 = nil/mild/moderate/severe) was used in four trials. Pain at rest and movement was assessed in nine trials and quantitative sensory testing was used in eight trials. A multidimensional score such as the McGill Pain Questionnaire was used in nine trials.

14Five of the trials studied the impact of the gabapentinoids and the dose regimen and duration of therapy was different in every trial (Fassoulaki et al.,,; Nikolajsen et al.,; Amr and Yousef,; Kinney et al., ). One trial incorporated the antidepressant venlafaxine (Amr and Yousef, ). Six trials studied the impact of ketamine; two consisted of intravenous ketamine without epidural analgesia (Hayes et al.,; Dualé et al., ), two consisted of intravenous ketamine in addition to epidural analgesia (Joseph et al.,; Mendola et al., ) and two examined the effect of epidural ketamine (Wilson et al.,; Ryu et al., ). Six trials studied the impact of cryoanalgesia in thoracotomy. Of these, two trials compared cryoanalgesia against control treatment with intravenous opioids (Gwak et al.,; Sepsas et al., ), one trial compared cryoanalgesia to epidural analgesia (Ju et al., ), two trials compared the effect of cryoanalgesia versus placebo in thoracotomy patients who all had epidural analgesia (Yang et al.,; Mustola et al., ) and one trial compared cryoanalgesia to non-divided intercostal muscle flap (Lu et al., ). Six trials examined the impact of local anaesthetic; of these, three incorporated infiltration with differing doses of ropivacaine in mastectomy (Albi-Feldzer et al.,; Baudry et al.,; Vigneau et al., ), one trial studied the impact of lidocaine infiltration in thoracotomy patients that already had a thoracic epidural in situ (Cerfolio et al., ), one trial examined the effect of intravenous lidocaine (Grigoras et al., ) and the remaining one assessed the impact of topical EMLA (eutectic mixture of local anaesthetic) cream in mastectomy patients (Fassoulaki et al., ). The impact of regional analgesia was studied in eight trials (two amputation, two mastectomy and four thoracotomy). Three trials examined whether commencing an epidural prior to thoracotomy was better than commencing it at the end of the procedure (Obata et al.,; Ochroch et al.,; Senturk et al., ). One trial compared pre-surgical epidural and low-dose intraoperative remifentanil infusion versus post-surgical epidural and high-dose intraoperative remifentanil infusion (Salengros et al., ). One trial compared epidural analgesia to epidural saline plus standard management of thoracotomy pain (Nikolajsen et al., ). Two trials examined the impact of paravertebral block for mastectomy patients (Kairaluoma et al.,; Ibarra et al., ). One trial had five study arms that studied epidural therapy, fentanyl patient-controlled analgesia (PCA) and standard therapy before, during and after amputation (Karanikolas et al., ). A single trial examined the impact of total intravenous anaesthesia (TIVA) versus inhalational anaesthesia in thoracotomy patients with thoracic epidurals in situ (Song et al., ).

15In patients after mastectomy, gabapentinoids reduced acute pain in three trials (Fassoulaki et al.,,; Amr and Yousef, ). Gabapentinoids also reduced the overall incidence of chronic pain in one (Fassoulaki et al., ) and the incidence of chronic burning pain in another (Fassoulaki et al., ). Gabapentin did not reduce acute pain in two trials where epidural analgesia was used for amputation or thoracotomy, respectively (Nikolajsen et al.,; Kinney et al., ). In the former, therapeutic doses of gabapentin were used, but the patients had chronic pain preoperatively secondary to vascular disease. In the latter, only a single dose of gabapentin was given perioperatively. The antidepressant venlafaxine (37.5 mg) was associated with significantly lower post-mastectomy pain after 7 days, and after 6 months, the incidence of post-mastectomy burning and stabbing pain was significantly lower (Amr and Yousef, ).

16In one trial, intravenous ketamine infusion significantly reduced acute pain after thoracotomy but did not reduce chronic pain (Dualé et al., ). In three trials, intravenous ketamine had no impact on acute or chronic pain with or without the presence of epidural analgesia for amputation or thoracotomy (Hayes et al.,; Joseph et al.,; Mendola et al., ). Epidural ketamine was co-administered with bupivacaine in two trials. In one of these, ketamine reduced acute pain but not chronic pain (Wilson et al., ), and in the other trial, acute pain was not assessed, but the drug had no impact on chronic pain (Ryu et al., ).

17In one trial of thoracotomy patients, intercostal cryoanalgesia reduced acute pain versus control therapy (no epidural; Sepsas et al., ). In the same trial, there was a difference in pain scores at 2 months that was statistically significant, but not clinically significant (VAS: 0.0/10 vs. 0.25/10). However, in a further five thoracotomy trials, cryoanalgesia had no impact on acute pain (Gwak et al.,; Yang et al.,; Ju et al.,; Mustola et al.,; Lu et al., ). Of these five thoracotomy trials, intercostal cryoanalgesia had no impact on chronic pain in one trial (Gwak et al., ) and actually increased chronic pain in four trials (Yang et al.,; Ju et al.,; Mustola et al.,; Lu et al., ).

18Six trials investigated the effect of local anaesthetics. Ropivacaine infiltration reduced acute pain for 1.5 and 6 h, respectively, in two trials (Vigneau et al.,; Albi-Feldzer et al., ), but had no impact on the development of chronic post-mastectomy pain. In a third ropivacaine trial, the drug had no impact on acute or chronic pain (Baudry et al., ). Lidocaine infiltration also had no impact on acute or chronic pain in thoracotomy patients (Cerfolio et al., ), whereas in one trial, intravenous lidocaine infusion (1.5 mg/kg pre- and intraoperative) had no impact on acute pain overall, but significantly reduced the incidence chronic pain and the area of hyperalgesia (Grigoras et al., ). In another trial, topical EMLA cream applied (for 5 days) had no impact on acute pain, but it reduced the incidence of chronic pain after mastectomy (Fassoulaki et al., ).

19Seven trials investigated the acute and chronic impact of regional anaesthesia including whether it could have a beneficial pre-emptive effect. One study found that although epidural analgesia was effective in the acute perioperative setting, it did not reduce chronic phantom limb pain in patients who had chronic pain preoperatively secondary to vascular disease (Nikolajsen et al., ). In contrast, two trials found that epidural analgesia reduced acute and chronic pain versus control therapy after thoracotomy and amputation, respectively (Senturk et al.,; Karanikolas et al., ). Four trials examined whether the timing of epidural analgesia could impact on acute and chronic pain (Obata et al.,; Ochroch et al.,; Senturk et al.,; Karanikolas et al., ). One of these trials found that it made no difference to acute or chronic pain whether the epidural was commenced immediately before or after the thoracotomy procedure (Ochroch et al., ). In contrast, the remaining three trials found that the optimization of analgesia preoperatively had a beneficial impact on both acute and chronic pain (Obata et al.,; Senturk et al.,; Karanikolas et al., ). Two trials of mastectomy patients investigated preoperative paravertebral block. The first trial found that paravertebral block significantly reduced acute pain and chronic pain 1 year later compared with sham block (Kairaluoma et al., ). The second trial found that paravertebral block had no significant impact on acute pain, but reduced the prevalence of chronic pain, compared with control (Ibarra et al., ). One trial compared preoperative epidural commencement with low-dose intravenous remifentanil infusion versus post-operative epidural commencement with a high-dose intravenous remifentanil infusion for thoracotomy patients (Salengros et al., ). They found no difference in acute pain, but high-dose remifentanil with a post-operative epidural was associated with a higher incidence of chronic pain. In addition, they found that the surface area of chest wall allodynia was significantly larger for the high-dose remifentanil group during the first post-operative month.

20A single large trial of thoracotomy patients receiving thoracic epidurals compared TIVA with propofol and remifentanil against inhalational anaesthesia with sevoflurane (Song et al., ). There was no difference in acute pain between the two groups, but the incidence of chronic pain was significantly reduced at 3- and 6-month assessments in the TIVA group.

21Gabapentinoids, antidepressants, local anaesthetics and regional anaesthesia have the potential to reduce the severity of both acute and chronic pain for patients undergoing amputation, thoracotomy or mastectomy. Ketamine and cryoanalgesia do not appear to be promising agents for reducing chronic pain. TIVA with propofol may have the potential to reduce the incidence of chronic pain, whereas the effect of opioids such as remifentanil is unclear.

22This systematic review was limited to amputation, mastectomy and thoracotomy because these operations are considered to be associated with a combination of nociceptive and neuropathic symptoms and carry a high risk of chronic pain. If it is accepted that the severity of acute pain is correlated with the risk of developing chronic pain, then it follows that the perioperative period is a logical target for interventions aimed at reducing or preventing the phenomenon. The positive and the negative results of this review are both important to guide future clinical practice. Specifically, it is useful to note the ineffective or harmful therapies so that they may be avoided and further trials of them may also be avoided.

23The trials included in this systematic review were of varying quality, but addressed the majority of conventional methodological standards. However, the heterogeneity of the trial methodology and outcome measures ruled out the performance of a meaningful meta-analysis. Despite the lack of consistency and comparability between trials, several consistent themes emerge in the results. Acute pain in the perioperative period after amputation, mastectomy and thoracotomy appears to be a mix of nociceptive and neuropathic pain. The acute neuropathic component is potentially responsive to agents that work on neuropathic symptoms, but not nociceptive symptoms. The acute neuropathic component may be relatively resistant to conventional analgesics, but is potentially treatable, and if the correct intervention is utilized in the correct manner (dose/duration/timing), it has the potential to reduce the development of chronic pain. A number of trials included in this review studied a single, brief intervention, which was typically not enough to achieve long-term benefits, potentially due to the fact that the process of neuroplasticity that can continue for days or weeks post-operatively. The days following an operation appear to be as important as the day of the operation itself, i.e., the potential benefits of a pain-free initial 12 h could be made obsolete if the pain is severe thereafter. Furthermore, neuropathic symptoms may take several days to manifest post-operatively, thus emphasizing the importance of recognition and referral to a vigilant and cognizant acute pain service.

24There were mixed results for the gabapentinoids; these agents reduced acute post-operative pain unless an effective epidural was present, in which case no significant difference was observed. Within the trials, gabapentinoids were often given once only or at sub-therapeutic doses. However, if an appropriate therapeutic dose was administered for a longer period of time, the efficacy appeared greater. This is in keeping with other studies of acute post-mastectomy pain reported in the literature (Dirks et al.,; Grover et al.,; Kim et al., ). Of note, vascular patients with chronic pain preoperatively continued to have chronic pain post-operatively despite gabapentin. Separately, the antidepressant venlafaxine was associated with enhanced analgesia 1 week after mastectomy and after 6 months despite being administered at the potentially sub-therapeutic dose of 37.5 mg.

25Interpretation of the trials examining the potential benefits of regional analgesia as a single group was challenging due to the inherent heterogeneity of trial methodology. Effective regional techniques in the hands of skilled clinicians can ablate the peak noxious barrage (Bonhomme et al.,; Wennervirta et al., ) of pain in the perioperative period and potentially minimize pathological neural plasticity implicated in the genesis of chronic pain (Woolf and Salter,; Scholz and Woolf,; Harvey and Dickenson,; Ivani et al., ). Unsurprisingly, the majority of the trials included in this review found regional analgesia to be beneficial for reducing perioperative pain and subsequent chronic symptoms. The single trial that found epidural analgesia to be no better than standard analgesia comprised vascular patients with pre-existing chronic pain prior to amputation. It is perhaps unsurprising that a few days of epidural analgesia did not impact on the chronic symptomatology of these vasculopaths. Several trials examined the timing of regional intervention. In particular, the complex trial by Karanikolas et al. indicated the benefits of early epidural insertion and continued infusion for the entire perioperative period. However, the same trial also found that if patients were managed aggressively and effectively with systemic opioids, an acceptable outcome could also be achieved. Early commencement of regional analgesia may offer some benefits for post-operative pain and may have other advantages such as facilitation of balanced anaesthesia, minimization of sympathetic stress on the cardiovascular system and potential reduction of blood loss. They are also particularly useful in patients who are opioid tolerant. Optimal duration of regional analgesia is yet to be determined and potential benefits may be lost if epidural infusions are discontinued too early for logistical reasons.

26Infiltration of local anaesthetic could provide a brief period of analgesia, but did not impact on the risk of developing chronic pain. In contrast, intraoperative intravenous lidocaine and perioperative EMLA cream reduced the incidence of chronic pain following mastectomy, without making a significant impact on acute pain.

27Remifentanil infusions have become a popular component of balanced anaesthesia for a multitude of surgical procedures. However, the theoretical concept of opioid-induced hyperalgesia appeared to be confirmed in a relevant clinical setting in the trial by Salengros et al.. Indeed, high-dose intraoperative remifentanil was also associated with an increased risk of chronic post-surgical pain. In contrast, TIVA with propofol and remifentanil was associated with a reduced incidence of chronic post-thoracotomy pain than inhalational anaesthesia with sevoflurane (Song et al., ). Given the phenomenon of opioid-induced hyperalgesia (Ballantyne and Mao,; Angst and Clark,; Raghavan et al., ), the potential benefit of TIVA is more likely to be propofol-mediated, potentially via glutamate, GABAA and glycine receptors (Song et al., ).

28While ketamine is an effective co-analgesic agent for acute pain and has the advantage that it may be used by patients in whom the oral route is unavailable, it did not reduce the risk of chronic post-surgical pain and its use in general is limited by dysphoric side effects. Ketamine also offered no additional benefit in the presence of effective epidural analgesia. Of even greater concern is intercostal nerve cryoanalgesia, which could arguably be renamed as ‘cryoneuralgia’ and perhaps be considered as an adjunctive therapy only in the setting of palliative care.

29Current mainstream perioperative management for amputation, thoracotomy and mastectomy is highly variable, but the choices that anaesthetists make may have long-term implications for patients under their care. Standard practices that rely primarily on opioids and NSAIDs for perioperative analgesia may be inadequate for patients at risk of neuropathic pain. Separately, techniques such as cryoanalgesia should be avoided. Gabapentinoids and antidepressants (which are not considered to be effective for nociceptive symptoms), as well as local anaesthetics and regional anaesthesia, may potentially reduce the severity of both acute and chronic pain for patients undergoing operations associated with a high risk of chronic post-surgical pain. However, some pharmacological agents may take a number of hours to be of benefit and therefore therapeutic dosages should be used for several weeks in many instances.

30There have been promising results for the use of gabapentinoids in other conditions (Brogly et al.,; Buvanendran et al.,; Gray et al., ) but anti-neuropathic pain drugs have not been proven to be beneficial for surgical interventions that are predominantly associated with nociceptive pain such as minor gynaecological procedures. Cases of neuropathic pain may be discovered using simple screening tools routinely, e.g., DN4 or LANSS (Bennett,; Bouhassira et al., ), and by identifying those who are experiencing a disproportionate level of pain. Appropriate interventions at this stage may provide superior analgesia and facilitate optimum recovery. Not all cases of chronic post-surgical neuropathic pain may be prevented, but the risk may be reduced. This may decrease the burden on society and the amount of financial resources spent on the management of chronic pain syndromes.

31Future trials in this field should focus on conditions with a high risk of neuropathic pain, e.g., amputation, thoracotomy and mastectomy. They should use appropriately high dosages of medications, and therapy could commence prior to surgery and continue for several weeks. Patient follow-up should be for a minimum of 6 months and a much more comprehensive and multidimensional assessment made of the character and severity of pain at rest and movement with an emphasis on level of function and basic neurophysiological testing (Melzack,; Wilder-Smith et al.,; Kehlet and Rathmell,; Wildgaard et al., ).