This article is accompanied by the following Invited Commentary:
Mion G. Ketamine stakes in 2018. Right doses, good choices. Eur J Anaesthesiol 2019; 36:1–3.
Postoperative pain management in chronic pain patients remains a clinical challenge. In the context of the current opioid epidemic, alternative nonopioid strategies are urgently needed to complement traditional opioid analgesia in the postoperative period. Spinal fusion patients often suffer from chronic pain and opioid tolerance, which makes postoperative pain control difficult. Opioid side effects such as nausea and sedation impede recovery. Chronic pain in spinal surgery patients is often of multifactorial origin.
There is evidence that opioid tolerance can occur in the dorsal root ganglion via N-methyl-D-aspartate (NMDA) receptor activation. 1,2 Increased synaptic transmission, or synaptic plasticity, from the dorsal horn neurones that carry the peripheral pain information to the spinal dorsal horn neurones can occur in the chronic pain state, and this form of plasticity has been referred to as central sensitisation. 3,4 Central sensitisation has been shown to be dependent on NMDA receptor activation. Both short-term and long-term administration of opioids cause acute opioid-induced hyperalgesia. 3 Ketamine, a noncompetitive NMDA receptor antagonist, has been hypothesised to counter opioid tolerance and NMDA receptor-mediated central sensitisation. 5,6 In addition, ketamine has been shown to function as an additive or complementary analgesic when used in combination with opioids in patients who display opioid tolerance. 7–9
Clinical studies have demonstrated reduced postoperative opioid requirements and improved pain control with subanaesthetic intra-operative ketamine infusions in opioid-tolerant patients. 8,9 In opioid-naïve patients, however, the benefits of intra-operative and postoperative use of ketamine were less convincing. 10 Therefore, clinical studies are needed to clarify the benefit of ketamine in postoperative pain control for opioid-tolerant versus opioid-naïve patients. Ketamine is currently not approved by the Food and Drug Administration to reduce opioid requirements.
The aims of our study were to examine whether postoperative ketamine infusion would reduce postoperative hydromorphone consumption in opioid-tolerant patients to a greater degree than it did in opioid-naïve patients, and whether ketamine would improve pain control in opioid-tolerant patients to a different degree than in opioid-naïve patients. The primary outcome was 24-h postoperative hydromorphone consumption and the secondary outcome was numerical pain scores for the first 24 h. Potential side effects were also recorded and analysed as additional secondary outcome measures.
This prospective, randomised, double-blind, placebo-controlled, four-arm parallel, single-centre study received ethical approval (IRB number S12-02202) from the New York University Institutional Review Board, New York, United States on 8 October 2012. The study was registered as NCT03274453 with clinicaltrials.gov.
The study was performed at the NYU Hospital for Joint Diseases, NYU Langone Medical Center, during the period from November 2012 to November 2014. All patients gave written informed consent before participating in the study.
Patients aged 16 to 75, ASA physical status 1 to 3, scheduled for elective lumbar fusion surgery at two or more levels under general anaesthesia, were prospectively studied. We defined opioid-tolerant as the daily use of opioid pain medication (oxycodone, morphine, hydromorphone, fentanyl, methadone or tramadol) during the 2 weeks before surgery. Patients who did not fulfil that criterion were deemed to be opioid-naïve. Exclusion criteria were poorly controlled hypertension, severe cardiac or pulmonary disease, elevated intraocular pressure, severe hepatic or renal dysfunction, pregnancy, a history of psychiatric disorder, inability to speak English, inability to understand the numerical pain scale or to operate the patient-controlled analgesia (PCA) pump, and known allergy to ketamine or hydromorphone.
Patients were allocated to the opioid-naïve or opioid-tolerant arms as defined above, and subsequently randomised into two groups, for a total of four groups: opioid-naïve ketamine, opioid-naïve placebo, opioid-tolerant ketamine and opioid-tolerant placebo. Patients in each group were randomised to receive either a ketamine infusion [ketamine bolus 0.2 mg kg−1 over 30 min, started on arrival in the postanaesthesia care unit (PACU), followed by a fixed-rate infusion of 0.12 mg kg−1 h−1 for 24 h], or placebo (identical volume/rate of 0.9% saline). Randomisation within each group was performed by the study coordinator using a computer-generated random number list in a 1 : 1 ratio. Study medication and placebo were produced according to the randomisation list in identical and consecutively numbered 250-ml bags. The pharmacist was not involved in patient care. Information about treatment was concealed but available for unblinding in case of acute complications. During the entire study period, investigators performing the postoperative assessments, medical staff (nurse, anaesthetist and surgeon) and subjects were blinded to group allocation.
In all patients, general anaesthesia was induced with propofol based on patient weight. Rocuronium 0.6 to 1.2 mg kg−1 was used to facilitate tracheal intubation. Anaesthesia was maintained with propofol (variable rate to maintain bispectral index at a level acceptable for surgical anaesthesia), desflurane less than 1.5% in a mixture of air and oxygen, and fentanyl, sufentanil, hydromorphone or morphine at the discretion of the anaesthetic staff. Blood pressure was maintained within 20% of baseline and hypotension was treated at the discretion of the anaesthetic staff with 0.9% sodium chloride solution, hetastarch, ephedrine and phenylephrine intravenously.
Most patients received 1 g of intravenous paracetamol at the end of surgery (Table 2). NSAIDs were not used in the immediate peri-operative period because of the surgeons’ concern that they might impede bone healing and fusion.
In all patients, postoperative pain treatment during the first 24 h consisted of standard care of intravenous PCA with hydromorphone (0.2-mg dose, lockout time 6 min, maximum rate 2 mg h−1), started on arrival in PACU. Pre-operatively, the patients were educated by the nursing staff in the use of the PCA pump and the numerical pain scale (NPS; 0 = no pain, 10 = worst imaginable pain). Rescue medication of intravenous hydromorphone 0.2 to 0.3 mg as needed was administered by a nurse with the goal to reduce the NPS to less than 4. Opioid pain medication was restricted to hydromorphone to allow for valid comparison between the groups. After 24 h, PCA was discontinued and all patients were treated according to the surgical department's standard regimen. Diazepam 2 mg was administered intravenously if needed for severe muscle spasms.
Moderate-to-severe nausea or vomiting was treated with intravenous ondansetron 4 mg. If ondansetron was ineffective, intravenous metoclopramide 10 mg was administered.
All postoperative assessments were performed by the study investigators or trained nurses blinded to group allocation. Cumulative hydromorphone consumption was calculated from 0 to 24 h postoperatively. NPS was recorded at arrival in PACU, every 30 min during the first 2 h, and then every 2 h on the ward during the first 24 h after surgery while patients were awake.
The primary outcome was cumulative hydromorphone consumption during the first 24 h after surgery. Secondary outcome was NPS in the same time period. We also recorded central nervous system adverse events during the same period.
In the study by Subramaniam et al., 10 the mean 24-h cumulative hydromorphone dose used by opioid-dependent patients undergoing major spinal surgery was 19.4 mg, with a SD of 13.6 mg. Based on this, 45 patients per group would be needed to detect a 30% reduction in cumulative hydromorphone use in the opioid-tolerant group treated with ketamine and at least 50% reduction in the opioid-naïve groups (as per Gottschalk et al. 11 ), with a power of 80% and α of 0.05. We planned to enrol 50 subjects per arm to allow for possible dropouts. An interim analysis using the O’Brien–Fleming stopping boundaries was planned after 50% of the patients in the opioid-tolerant group were enrolled. 12
We tested normality using the Shapiro–Wilk test and Q–Q plots. Continuous data were compared using one-way analysis of variance (ANOVA) or the Kruskal–Wallis H test, as appropriate. The χ2 test and Fisher's exact test were used for inferences on proportions. A linear mixed model was used to analyse the effect of ketamine on cumulative hydromorphone consumption over time between groups, to accommodate an unbalanced design (unequal number of subject per group), and missing data in the repeated-measures design (missing at random). The covariance structure was chosen based on the Akaike information criteria and the Bayesian information criteria. In the presence of a significant interaction term or main effect, post hoc analyses using Bonferroni correction for multiple comparisons was used. Since postoperative pain scores did not follow a multivariate normal distribution, the analysis was performed using a rank-based repeated measures ANOVA adjusted for baseline score (immediately before starting the placebo/ketamine infusion). 13
Data are expressed as mean ± SEM or median [IQR], unless stated otherwise. A two-sided P value of less than 0.05 was considered significant. All analyses were performed with STATA/SE version 14.1 (StataCorp LP, College Station, Texas, USA) and R software version 3.3.2 (R Foundation, Vienna, Austria), according to the intention-to-treat principle.
A total of 249 patients were assessed for eligibility from November 2012 to November 2014. Of these, 129 patients were prospectively allocated randomly to groups (Fig. 1), but only 122 received the intended study intervention. Of those patients who received the study drug, 11 were excluded from the analysis due to incomplete clinical data (data inaccessible after new Electronic Health Record system implementation).
The study was stopped after the interim analysis of the primary outcome variable showed a highly statistically significant effect on the 24-h cumulative hydromorphone use between groups (n=111, P < 0.001; O’Brien–Fleming efficacy boundary P = 0.005).
Baseline characteristics and intra-operative data are shown in Tables 1 and 2, respectively. One patient in each group had only one level fused despite having been initially scheduled for two. Opioid-tolerant placebo patients had a higher BMI than those in the opioid-tolerant ketamine group (post hoc analysis using Dunn's pairwise comparison). Differences in pre-operative pain scores and opioid use, using morphine equivalents, were significant, as expected, between opioid-naïve and opioid-tolerant groups. When comparing only the two opioid-tolerant groups, the pain scores were significantly different, with the tolerant-placebo group reporting higher scores, but the difference in pre-operative opioid use did not reach statistical significance.
The time course of the postoperative cumulative hydromorphone consumption is shown in Fig. 2. Based on the analysis, the ‘group’ × ‘time’ interaction was statistically significant (P < 0.001), indicating that the effect of ‘group’ was not the same across all levels of ‘time’ (repeated measures). The contrast analysis of ‘group’ × ‘time’ (effects for group at each level of time) showed that the cumulative hydromorphone use (per kg) between the opioid-tolerant ketamine and opioid-tolerant placebo group became significantly different after 16 h of ketamine infusion (Bonferroni corrected P = 0.003), a difference that persisted until the end of the study. No difference was seen between the opioid-naïve ketamine and placebo groups, or between the opioid-tolerant ketamine and opioid-naïve placebo groups.
Moreover, treating ‘time’ as a continuous variable, the rate of change (slope) in the cumulative hydromorphone use per hour was calculated. The rate of hydromorphone use was 0.004 (95% CI 0.003 to 0.005) and 0.005 (95% CI 0.004 to 0.006) mg kg−1 h−1 in the opioid-naïve ketamine and placebo group, respectively (Bonferroni corrected P = 0.118). The opioid-tolerant ketamine group used hydromorphone at a rate of 0.007 (95% CI 0.006 to 0.008) mg kg−1 h−1, which was significantly less than the rate used by the opioid-tolerant placebo group [0.011 mg kg−1 h−1 (95% CI 0.010 to 0.011), Bonferroni corrected P < 0.001].
Figure 3 shows the postoperative pain scores in each group. The analysis, adjusted for baseline pain score, showed that only the main effect of ‘group’ was statistically significant (‘group’ P < 0.001, ‘time’ P = 0.495). The post hoc analysis using Bonferroni correction for the main effect ‘group’ showed that the pain scores in the opioid-tolerant placebo group were always higher in magnitude compared with the other three groups, but the study intervention did not affect the postoperative pain profile (‘group’ × ‘time’ interaction P = 0.192). There were no significant differences in side effects among groups (Table 3).
We found that postoperative ketamine infusion significantly reduces the hydromorphone requirement specifically for the opioid-tolerant patients after spinal surgery. The opioid-tolerant ketamine group used hydromorphone at a rate of 0.007 mg kg−1 h−1, a rate significantly lower than that used by the opioid-tolerant placebo group (0.011 mg kg−1 h−1) suggesting a strong opioid sparing effect. However, the hydromorphone consumption of the opioid-naïve group who were randomised to receive ketamine was not statistically different from that of the opioid-naïve group who received placebo. Hydromorphone requirement in the opioid-tolerant group on ketamine did not differ significantly from that in both opioid-naïve groups. The postoperative ketamine administration seemed to readjust pain medication requirements of opioid-tolerant patients to a range comparable with that of opioid-naïve patients. The interventions did not affect the pain score behaviour among groups, even though pain scores were always highest in the opioid-tolerant placebo group. One possible reason for a lack of difference in pain scores may be that patients on PCA self-adjust to a comfortable level of pain.
Although opioid-related side effects did not differ among groups, we still believe that these findings are interesting. This study was conceived as a proof of concept, to demonstrate the efficacy of ketamine in reducing opioid requirements in opioid-tolerant patients. Indeed, reducing the dose of opioid taken without reducing side effects might not seem like a worthy outcome, but it might lead to less opioid-related pain sensitisation, lower long-term opioid use and lower rates of chronic pain and opioid addiction.
Intra-operative ketamine administration for postoperative pain has been shown to be beneficial in opioid-tolerant patients. 8,9 In a review of clinical trials and meta-analyses, Jouguelet-Lacoste et al. 14 showed a reduction in opioid consumption by 40%. The primary endpoint of the majority of the studies included in this meta-analysis was opioid consumption (24 out of 39 studies), whereas the secondary endpoint was most frequently reduction in pain scores (10 out of 39 studies). Our review of the literature demonstrated several difficulties in evaluating the usefulness of ketamine as an adjunct to postoperative pain medication, for example heterogeneity in operative procedures (intensity of pain depending on the surgery, somatic versus visceral pain), in patient populations (chronic pain patients versus patients who do not suffer from chronic pain), in the divergent modes of ketamine administration (bolus versus infusion, intra-operative versus postoperative versus a combination of intra-operative and postoperative infusion) and finally in differences in concomitantly administered pain medications (including both opioids and nonopioids). 10,14–16 Our study was designed to address some of these difficulties by recruiting from a relatively homogeneous surgical population and using a standardised analgesic protocol. We chose not to use ketamine intra-operatively as we feared that it would lead to very different doses of intra-operative opioids used, and thus muddle the result. Also, surgical procedures of different length could have led to vastly different total doses of ketamine being administered.
We hypothesised that several mechanisms in the generation of postoperative pain determine how well opioids and ketamine work in the treatment of this pain. Chronic pain leads to changes in pain processing. These changes manifest as peripheral and spinal central sensitisation, as well as disrupted cortical and subcortical processes in the brain. 17–21 Furthermore, chronic pain is known to be associated with a host of psychosocial comorbidities such as anxiety and depression. 2,22 Opioids are usually prescribed for chronic pain. Chronic opioid use can result in tolerance and dependence 2 as well as hyperalgesia. 1 Even acute opioid administration can alter spinal and peripheral nociceptive processing, resulting in opioid-induced hyperalgesia within a short time period. 3,4,23 These mechanisms of tolerance and sensitisation can rapidly reduce the opioid benefit for pain control. Central sensitisation and tolerance, meanwhile, are in part mediated in the dorsal horn neurones as well as in the cortex by NMDA receptor signalling, and NMDA receptors can be antagonised by ketamine. 24–27 More recently, ketamine has also been shown to improve mood, and it is currently being investigated for use in depression. The postoperative period can be accompanied by a depressed mood, and ketamine has been shown in animal studies to improve mood in the postoperative setting. 28 One target for its mood-enhancing properties lies in the prefrontal cortex. 29 Thus, ketamine may also be able to relieve the affective component of pain independent from its modulation of the nociceptive transmission pathway. Our opioid-tolerant chronic pain patients did benefit substantially more from postoperative low-dose ketamine infusion than did opioid-naïve patients who had no significant pre-operative pain. There are several possible reasons for this selective effect of ketamine analgesia. First, ketamine could antagonise NMDA receptors in the spine and brain to decrease opioid tolerance and chronic pain. Second, ketamine infusion could provide additional relief of the aversive component of pain as well as an improvement in mood via its activity in the prefrontal cortex and associated areas.
It is important to differentiate between benefits from ketamine because of its action in reversing the effects of chronic pain and opioid tolerance on one hand and its benefits due to a preventive effect against acute opioid induced hyperalgesia on the other hand. Two other studies have examined the use of ketamine in opioid-tolerant patients. Loftus et al. 8 studied effects of intra-operative ketamine administration exclusively in opioid-tolerant patients and found a 37% reduction in opioid consumption during the initial 48-h postoperative period and a 71% reduction at 6 weeks after surgery. They also found reductions in pain scores 48 h and 6 weeks after surgery. The authors hypothesise that the improved pain control at 6 weeks was due to a reduction of central sensitisation due to NMDA receptor antagonism and to improved acute pain control in the immediate postoperative period. Problems with this study included a lack of standardisation in the opioid regimen: different opioid classes were administered. Nielsen et al. 9 found comparable results in opioid-tolerant patients. Both groups also noticed an increased benefit for patients who had been on substantial amounts of pre-operative opioids, over 40 mg daily in the Loftus study and over 36-mg morphine equivalent daily in the Nielsen study. Unlike these authors, we administered ketamine in the postoperative period, but our results show a similar reduction in opioid requirements. Our data suggest that a successful analgesic intervention in chronic-pain/opioid-tolerant patients does not necessarily need to be implemented prior to the surgical insult.
A problem which we encountered in our patient selection was that our opioid-naïve patients more frequently were younger and did not suffer from chronic pain because they presented for surgery for scoliosis correction, while the opioid-tolerant patients had often suffered from chronic back pain with concomitant psychological effects and expectations of pain medication. Another shortcoming of our study is that we did not precisely quantify the amount of pre-operative opioid medication taken by the patients, so we are unable to confirm Loftus’ and Nielsen's observations that the benefit of ketamine increases with the amount of pre-operatively administered opioids. Finally, our study was not designed for a follow-up over a long time period. It would be interesting to determine whether opioid-naïve patients derive any benefit from ketamine in the longer postoperative term, such as a reduction in the risk of developing chronic pain.
In conclusion, our study shows that low-dose ketamine infusion postoperatively reduces postoperative opioid requirements and pain in opioid-tolerant patients who used opioids pre-operatively, but that it is of less benefit for opioid-naïve patients in the immediate postoperative period. It is unclear whether other adjuvants, such as gabapentinoids, would have a similar effect or possibly potentiate the effect of ketamine. In the context of the current opioid epidemic, such investigations on nonopioid analgesic strategies will have a large impact on postoperative care and public health.
Acknowledgements relating to this article
Assistance with the study: we thank James McKeever, Emily Siu and Randy Cuevas.
Financial support and sponsorship: this work was supported by the Department of Anesthesiology, Perioperative Care and Pain Medicine, New York University School of Medicine, New York, NY, USA.
Conflicts of interest: AA has received consulting fees from Pacira and Trevena and speaker fees from B Braun.
Presentation: preliminary data from this study were presented in part at the International Anesthesia Research Society annual meeting, 17 to 20 May 2014, Montreal, Canada, and at the European Society of Anaesthesiology Euroanaesthesia meeting, 30 May to 2 June 2015, Berlin, Germany.
1. Mao J. Opioid-induced abnormal pain sensitivity: implications in clinical opioid therapy. Pain
2. Mayer DJ, Mao J, Holt J, et al. Cellular mechanisms of neuropathic pain, morphine tolerance, and their interactions. Proc Natl Acad Sci U S A
3. Simonnet G. Opioids: from analgesia to antihyperalgesia? Pain
4. Simonnet G, Rivat C. Opioid-induced hyperalgesia: abnormal or normal pain? NeuroReport
5. Rogawski MA. Therapeutic potential of excitatory amino acid antagonists: channel blockers and 2,3-benzodiazepines. Trends Pharmacol Sci
6. Shimoyama N, Shimoyama M, Inturrisi CE, et al. Ketamine attenuates and reverses morphine tolerance in rodents. Anesthesiology
7. Ding X, Jin S, Niu X, et al. Morphine with adjuvant ketamine versus higher dose of morphine alone for acute pain: a meta-analysis. Int J Clin Exp Med
8. Loftus RW, Yeager MP, Clark JA, et al. Intraoperative ketamine reduces perioperative opiate consumption in opiate-dependent patients with chronic back pain undergoing back surgery. Anesthesiology
9. Nielsen RV, Fomsgaard JS, Siegel H, et al. Intraoperative ketamine reduces immediate postoperative opioid consumption after spinal fusion surgery in chronic pain patients with opioid dependency: a randomized, blinded trial. Pain
10. Subramaniam K, Akhouri V, Glazer PA, et al. Intra- and postoperative very low dose intravenous ketamine infusion does not increase pain relief after major spine surgery in patients with preoperative narcotic analgesic intake. Pain Med
11. Gottschalk A, Durieux ME, Nemergut EC. Intraoperative methadone improves postoperative pain control in patients undergoing complex spine surgery. Anesth Analg
12. Schulz KF, Grimes DA. Multiplicity in randomised trials II: subgroup and interim analyses. Lancet
13. Brunner E, Domhof S, Langer F. Nonparametric analysis of longitudinal data in factorial experiments. New York, USA: J. Wiley; 2002.
14. Jouguelet-Lacoste J, La Colla L, Schilling D, et al. The use of intravenous infusion or single dose of low-dose ketamine for postoperative analgesia: a review of the current literature. Pain Med
15. Bell RF, Dahl JB, Moore RA, et al. Peri-operative ketamine for acute postoperative pain: a quantitative and qualitative systematic review (Cochrane review). Acta Anaesthesiol Scand
16. Subramaniam K, Subramaniam B, Steinbrook RA. Ketamine as adjuvant analgesic to opioids: a quantitative and qualitative systematic review. Anesth Analg
17. Basbaum AI, Bautista DM, Scherrer G, et al. Cellular and molecular mechanisms of pain. Cell
18. Ji G, Zhang W, Mahimainathan L, et al. 5-HT2C receptor knockdown in the amygdala inhibits neuropathic-pain-related plasticity and behaviors. J Neurosci
19. Latremoliere A, Woolf CJ. Central sensitization: a generator of pain hypersensitivity by central neural plasticity. J Pain
20. Woolf CJ, Costigan M. Transcriptional and posttranslational plasticity and the generation of inflammatory pain. Proc Natl Acad Sci U S A
21. Zhang Q, Manders T, Tong AP, et al. Chronic pain induces generalized enhancement of aversion. eLife
22. Doan L, Manders T, Wang J. Neuroplasticity underlying the comorbidity of pain and depression. Neural Plast
23. Hayhurst CJ, Durieux ME. Differential opioid tolerance and opioid-induced hyperalgesia: a clinical reality. Anesthesiology
24. Coderre TJ, Melzack R. The role of NMDA receptor-operated calcium channels in persistent nociception after formalin-induced tissue injury. J Neurosci
25. Eide PK, Jorum E, Stubhaug A, et al. Relief of postherpetic neuralgia with the N
-methyl-D-aspartic acid receptor antagonist ketamine: a double-blind, cross-over comparison with morphine and placebo. Pain
26. Eide PK, Stubhaug A, Stenehjem AE. Central dysesthesia pain after traumatic spinal cord injury is dependent on N
-methyl-D-aspartate receptor activation. Neurosurgery
27. Hocking G, Cousins MJ. Ketamine in chronic pain management: an evidence-based review. Anesth Analg
28. Wang J, Goffer Y, Xu D, et al. A single subanesthetic dose of ketamine relieves depression-like behaviors induced by neuropathic pain in rats. Anesthesiology
29. Li N, Lee B, Liu RJ, et al. mTOR-dependent synapse formation underlies the rapid antidepressant effects of NMDA antagonists. Science