Adults presenting for deformity correction surgery have both inflammatory and neuropathic pain related to discogenic disease, facet arthropathy, ligamental hypertrophy, and ultimately segmental instability with the progressive development of nerve root and cord compression. Fu et al.1 reported a 97% and 44% prevalence of neuroforaminal and severe canal stenosis, respectively, in adults with degenerative scoliosis, which results in significant functional disability and pain. As a result of this, 16% to 33% of patients require chronic opioid therapy before back surgery.2
This combination of preexisting somatic and neuropathic pain with superimposed surgical pain and chronic opioid use can pose significant challenges for postoperative pain management. Targeting multiple antinociceptive and antihyperalgesic pathways can decrease opioid consumption, potentially reducing the incidence of opioid-related complications.
Non-opioid analgesic drugs with established opioid-sparing effects include N-methyl-D-aspartate (NMDA) antagonists, gabapentenoids, IV lidocaine, and α-2 agonists.3 α-2 Agonists demonstrate novel nociceptive-modulating effects through both central and spinal cord α-2 receptor binding and are independent of the route of administration, with epidural, intrathecal, and systemic administration all demonstrating an opioid-sparing effect.4
In spine surgery, previous studies demonstrating an opioid-sparing effect with α-2 agonists were confined predominately to lumbar laminectomy cases only.5,6 The role of highly selective α-2 agonists, such as dexmedetomidine on postoperative opioid consumption and pain scores with larger, more invasive spine procedures requiring instrumentation has not been studied.
Therefore, we tested the hypothesis that intraoperative dexmedetomidine reduces postoperative opioid consumption and improves pain scores in patients undergoing multilevel deformity correction spine surgery.
This was a prospective double-blinded, placebo-controlled, parallel group trial. The study was approved by the University of Virginia IRB (UVA HSR #16150) and was registered under Clinical Trial number: NCT01850017.
All participants undergoing multilevel spine surgery requiring neurologic monitoring were screened. We recruited participants aged between 18 and 80 years who were scheduled for multilevel (>3 levels) thoracic and/or lumbar spine surgery requiring either somatosensory evoked potential, motor evoked potential, and/or electromyography monitoring. Exclusion criteria included emergency spine surgery, pregnancy, advanced heart block (Mobitz type II or atrioventricular dissociation) on preoperative electrocardiogram, stage IV or greater renal disease (estimated glomerular filtration rate <30 mL/min, and/or requiring dialysis), liver failure (history of cirrhosis or fulminant hepatic failure), preoperative methadone use, or preoperative systolic hypertension (systolic blood pressure >150 mm Hg) immediately before surgery. Participants were enrolled from September 2012 to July 2015.
After informed written consent was obtained, a computer-generated block randomization scheme was used to stratify participants into 1 of 2 treatment regimens: dexmedetomidine or placebo (saline). Primary investigators, patients, anesthesia providers, and postoperative medical staff were blinded to the treatment assignment.
Before induction of anesthesia, participants were premedicated with 1 to 2 mg midazolam. Standard ASA monitors, an arterial line, and Bispectral Index monitor (BISTM, Covidien, Minneapolis, Minnesota) were used in all patients. Anesthetic induction was achieved with 1 to 3 mg/kg propofol, 0.5 to 1.5 mg/kg lidocaine, and 0.5–3 μg/kg fentanyl. Neuromuscular blockade was achieved with either 0.6 to 1.2 mg/kg rocuronium or 0.8–1.2 mg/kg succinylcholine. All weight-based dosing was calculated on ideal body weight (IBW).
After induction, general anesthesia was maintained with desflurane while additional monitoring lines were placed. Both groups received 0.2 mg/kg (IBW) of methadone IV for analgesia.
After transitioning to the prone position, desflurane was discontinued, and anesthesia was maintained using a total IV technique with propofol to facilitate neurophysiologic monitoring. Propofol was titrated to maintain BIS between 30 and 60. An initial loading dose of 1 μg/kg (IBW) dexmedetomidine was administered over 20 minutes followed by a maintenance infusion of 0.5 μg/kg/h (IBW). An equivalent volume of a sodium chloride solution (0.9%) was administered for the loading dose and infusion in the placebo group. Both dexmedetomidine and placebo were stopped when closure of the surgical wound began.
A 20% increase in heart rate and/or arterial blood pressure from the preoperative baseline (obtained in the surgical admission suite before surgery) or patient movement was treated with fentanyl boluses (0.5–1.5 μg/kg) at 2.5-minute intervals until vital signs returned to baseline or movement had stopped.7
Intraoperative hypotension was defined as a systolic and/or mean arterial blood pressure <100 and 70 mm Hg, respectively. Hypotension was initially managed with a combination of fluids, ephedrine, and/or phenylephrine bolus at the discretion of the anesthesiology team. If hypotension persisted despite the aforementioned interventions, phenylephrine infusion was commenced at 20 μg/min. The infusion rate was increased every 10 minutes by 20 μg/min to achieve the systolic and mean arterial blood pressure goals. The phenylephrine infusion was not increased beyond 100 μg/min. Persistent hypotension unresponsive to phenylephrine was treated with norepinephrine. The norepinephrine infusion was started at 2 μg/min and was increased by 1 μg/min every 5 minutes to a maximum of 10 μg/min until the blood pressure goals were achieved. Failure to reach and maintain the desired blood pressure goals triggered a reduction in the study drug to 0.35 μg/kg/h. Thirty minutes of persistent hypotension despite norepinephrine required that the study drug be stopped.
Intraoperative bradycardia, defined as heart rate <50/min, was initially treated with IV 0.1 mg glycopyrrolate. Glycopyrrolate was repeated at 2- to 3-minute intervals to a maximum of 0.4 mg. Bradycardia refractory to glycopyrrolate was treated with 0.4 to 0.8 mg atropine. If bradycardia persisted despite glycopyrrolate and atropine, study drug was reduced to 0.35 μg/kg/h. Ongoing bradycardia lasting >30 minutes at the reduced infusion rate triggered termination of the study infusion. Intraoperative antifibrinolytic therapy and blood product management was administered at the discretion of the anesthesia providers.
Postoperatively all study participants were admitted to the intensive care unit (ICU) consistent with our institutional practice. Postoperative opioid analgesia was administered by a standardized protocol in the ICU. This included the use of a hydromorphone patient-controlled analgesia pump with a patient-controlled bolus dose of 0.1 mg with a lockout period of 8 minutes. The ICU staff could adjust the bolus dose and lockout period if pain control was inadequate based on our standardized institutional practice. In addition, IV fentanyl could be administered before commencement of the patient-controlled analgesia pump and for breakthrough pain not adequately controlled by the hydromorphone pump. In addition to the IV opioid regimen, oral oxycodone at a dose of 5 mg every 4 hours was initiated when oral intake was commenced and increased if pain control remained inadequate. If participants had an allergy to oxycodone, oral hydromorphone or hydrocodone was substituted. All preoperative opioids, muscle relaxants, and gabapentenoids were restarted postoperatively. All participants received scheduled acetaminophen unless contraindicated. Nonsteroidal anti-inflammatory drugs were excluded based on our institutional practice.
Demographic data (age, sex, weight, height, and ASA classification), preoperative daily opioid consumption, non-opioid adjuncts (acetaminophen, gabapentin, pregabalin, baclofen, and cyclobenzaprine), duration of surgery, estimated blood loss, procedure type (number of levels, with or without instrumentation), methadone dose, study drug dose and infusion time, estimated blood loss, fluids, packed red blood cells, other blood products, autologous blood volume administered, inotropes, vasopressors, anticholinergics, BIS, heart rate, and blood pressure were recorded.
Postoperative data included time to tracheal extubation, incidence of hypotension defined as mean arterial blood pressure <60 mm Hg, vasopressor or inotropes, respiratory depression defined by a respiratory rate <8 breaths per minute, respiratory arrest, naloxone requirement, hypoxemia/desaturation defined as a SaO2 < 90% or the need for supplemental oxygen to maintain SaO2 >96% after 12 hours of extubation, self-reported postoperative nausea, documented vomiting, and delirium determined by the Confusion Assessment Method for ICU.
Spine Surgery Invasiveness Index
The Spine Surgery Invasiveness Index (SSII), previously validated by Mirza et al.,8 incorporates the number of levels decompressed, instrumented, and fused. Each operated vertebra can be assigned a score between 0 and 6 based on the procedural elements described and whether an anterior and/or posterior approach was performed. A summative score was calculated that corresponded to the extent of the procedure performed.
Opioid Consumption and Pain Scores
All opioids used during the intraoperative and postoperative period were converted to morphine equivalents for analysis (http://www.uptodate.com/contents/cancer-pain-management-with-opioids-optimizing-analgesia). Daily cumulative opioid consumption was calculated. A 11-point visual analog scale (VAS) was used to rate the postoperative pain. Participants who remained tracheally intubated postoperatively were not assigned a VAS score until they could participate in the evaluation.
The study was designed to target a 30% reduction in median opioid requirements for postoperative patients during the period from 24 to 48 hours after extubation. The study was designed with an overall 2-sided type I error rate of α = 0.05 and power 1 − β = 0.8. The study was designed with an interim analysis to be performed after enrollment of half the patients, using the study design by Pocock.9
Postoperative opioid requirements vary dramatically among patients, and our preliminary original data were highly skewed. However, on a log-scale, the data were approximately normally distributed. Thus, we proposed that the data be analyzed using the 2 sample t test after taking the natural log of each observation. The log transformation transforms a 30% reduction in median opioid requirements into a ln (1−0.3) = −0.356 change in mean opioid requirements. Thus, our study was powered to detect a −0.356 unit change in mean after transforming the data. The total number of participants required was 212 with the interim analysis after 53 participants were enrolled in each treatment group. After each batch of patients, the data were analyzed using a 2-sided 2-sample t test at level α = 2 × 0.0147 = 0.0294, which controls the overall α level at 0.05 after accounting for the interim and final analyses. If at the interim testing point the t test was significant at this α level, the study would have been terminated.
Categorical data were analyzed using the χ2 test, unless the expected number of events under the null hypotheses decreased below 5 for any combination of study drug and outcome; in these cases, Fisher exact test was used.
Continuous data were analyzed using the Wilcoxon rank sum test (equivalent to the Mann-Whitney U test), which is robust to the skewness exhibited in many of the continuous variables. Opioid requirements were also analyzed using the Wilcoxon test. A P value <0.05 was considered significant. All statistical analyses were conducted in R (version 3.2.2).
One thousand eight hundred fifty-two patients undergoing spine surgery with neurologic monitoring were screened for eligibility between 2012 and 2015. A CONSORT flow diagram is presented in Figure 1 highlighting the number of screened patients excluded because of the predefined criteria. One hundred forty participants undergoing >3 levels of thoracic and/or lumbar spine surgery requiring neurologic monitoring gave consent to participate in the study with 131 randomly assigned to 1 of the 2 treatment groups.
At the interim analysis, the 2 treatment groups showed almost identical opioid use after surgery (opioid use on postoperative day 2 produced a P value of 0.90), and a futility study was performed to assess the benefit of continuation. At this point, it was determined that if the effect size were as large as initially hypothesized, a continued study would have only 9% power. Enrollment continued for an additional 6 months until implementation of a standardized surgical and anesthesia practice protocol (Enhanced Recovery After Surgery) for major spine surgery at our institution, at which point a decision was made to terminate the study before complete enrollment. This was because of the limited power of the study to demonstrate a significant reduction in opioid use with dexmedetomidine, and the new institutional-approved standardized enhanced recovery program.
There were no significant differences in age, weight, ASA status, previous back surgery, and non-opioid medication use (acetaminophen, gabapentin, pregabalin, baclofen, and cyclobenzaprine) between the 2 groups (Table 1). Preoperative pain scores (placebo: 5 [1–7] versus dexmedetomidine: 4 [2–6], P = 0.4), percentage of participants with preoperative opioid use (placebo: 68% versus dexmedetomidine: 71%, P = 0.78), and daily opioid consumption between the groups (placebo: 7.5 [0–15] versus dexmedetomidine: 6.7 [0–17] mg morphine equivalents, P = 0.65) were not statistically different. Preoperative hemodynamic data (blood pressure and heart rate), respiratory rate, and oxygen saturation were comparable between the 2 groups.
The SSII, duration of surgery, propofol dose, crystalloid, colloid volume, and blood product utilization were similar between the treatment arms (Table 2). The significant intraoperative differences noted were (1) reduced opioid consumption in the dexmedetomidine group (placebo: 7 [3–15] versus dexmedetomidine: 3.5 [0–11] mg morphine equivalents, P = 0.04) and (2) increased incidence of bradycardia (placebo: 37% versus dexmedetomidine: 59%, P = 0.02), slower median heart rate (placebo: 72 [64–81] versus dexmedetomidine: 64 [60–70] per minute, P < 0.0001), and increased phenylephrine use in the dexmedetomidine group (placebo: 59% versus dexmedetomidine: 78%, P = 0.03).
No significant differences in opioid consumption were noted postoperatively (placebo versus dexmedetomidine, median [25%–75% interquartile range]: 24 hours (49 [30–78] vs 61 [34–77] mg morphine equivalents, P = 0.65), 48 hours (41 [28–68] vs 40 [23–64] mg morphine equivalents, P = 0.60), or 72 hours (29 [15–59] vs 30 [14–46] mg morphine equivalents, P = 0.58; Fig. 2). The Wilcoxon-Mann-Whitney odds is 1.11 with 97.06% confidence interval (0.71–1.76) for opioid consumption.10,11
Furthermore, no difference in pain score, measured by the 11-point VAS, was observed (placebo versus dexmedetomidine, median [25%–75% interquartile range]: 24 hours: 7 [5–7] vs 6 [4–7], P = 0.12 and 48 hours: 5 [3–7] vs 5 [3–6], P = 0.65; Fig. 3). Immediate postoperative VAS scores at 2 hours (placebo: 8 [6–10] vs dexmedetomidine: 7 [4–9], P = 0.28), 6 hours (placebo: 7 [5–8] versus dexmedetomidine: 6 [4–8], P = 0.25), and 12 hours (placebo: 7 [5–8] versus dexmedetomidine: 6 [4–8], P = 0.33) after surgery were not significantly different between the treatment arms. The incidence of nausea was similar between the 2 groups; however, there was an increased incidence of vomiting on postoperative day 1 in the dexmedetomidine group (placebo: 1.5% versus dexmedetomidine: 12.9%, P = 0.01) and day 3 (placebo: 0% versus dexmedetomidine: 6.5%, P = 0.05). No differences in the incidence of respiratory depression, naloxone use, vasopressor requirements, or delirium were observed in the postoperative period (Table 3).
This study demonstrates that intraoperative administration of dexmedetomidine is not associated with reduced postoperative opioid consumption or improved pain scores up to 3 days after major reconstructive spine surgery. Furthermore, immediate postoperative VAS scores at 2, 6, and 12 hours were no different between the 2 groups. There was a modest but significant reduction in intraoperative opioid administration in the dexmedetomidine group. This effect may have been related to the reduced frequency of blood pressure and heart rate-triggered opioid dosing in our study because of the sympatholytic effects of dexmedetomidine. However, a concomitant antinociceptive effect of dexmedetomidine cannot be excluded. Dexmedetomidine was associated with an increased incidence of bradycardia, slower median heart rate, and an increased need for phenylephrine intraoperatively. These hemodynamic changes were not sustained in the postoperative period. Finally, our study demonstrated a slightly increased incidence of vomiting in the dexmedetomidine group on postoperative days 1 and 3.
Our findings are in contrast to previous studies that have reported reduced opioid use in the immediate postoperative period and up to 48 hours after surgery.12–15 In a meta-analysis of 30 studies with 1792 patients, Blaudszun et al.16 demonstrated reduced opioid consumption at 24 hours with both clonidine (weighted mean difference was −4.1 mg [95% confidence interval, −6.0 to −2.2]) and dexmedetomidine (weighted mean difference −14.5 mg [95% confidence interval, −22.1 to −6.8]). Dexmedetomidine also reduced pain intensity (weighted mean difference, −0.6 cm) compared with the control group (3.1 cm [range, 1.9–4.0]) 24 hours after surgery.
A possible reason for the lack of a long-term effect of dexmedetomidine in our study was the use of methadone for intraoperative analgesia. Methadone is a µ-opioid agonist and NMDA antagonist with a long half-life that makes it ideally suited for the management of acute, chronic, and neuropathic pain. We have previously shown that in spine surgery, a single dose of intraoperative methadone reduces opioid consumption by 50% at 48 hours compared with an intraoperative sufentanil infusion.17 Interestingly, in 2 other prospective randomized placebo-controlled trials evaluating lidocaine and ketamine for multilevel spine surgery, intraoperative morphine equivalent consumption in the treatment arm was substantially higher than in our cohort (lidocaine: 36 [23–60] mg and ketamine: [51 ± 27 mg]).2,18 This suggests that a single dose of methadone can provide comprehensive intraoperative pain control with sustained opioid-sparing effects because of its prolonged half-life. Thus, methadone can possibly mask the salutary benefits of dexmedetomidine. Adding lidocaine or ketamine to our study would have made it even less likely that a beneficial effect of dexmedetomidine would have been identified.
Another possible reason why our findings are different may be related to the surgical procedure. The median SSII in our study for the placebo and dexmedetomidine arms were 14 (8–20) and 12 (8–22), respectively, indicative of multiple levels of instrumentation and osteotomies with significant tissue injury and inflammation. This is in contrast to procedures where an α-2 agonist has shown benefit, namely abdominal, intracranial, and minor spine surgery. Ozkose et al.6 reported a better hemodynamic profile and lower pain scores with intraoperative dexmedetomidine, whereas the addition of dexmedetomidine to a postoperative morphine patient-controlled analgesia pump reduced opioid use with slightly increased sedation in the study by Gunes et al.5 The major difference in both aforementioned studies was that surgery was limited to lumbar laminectomy cases only.
Interestingly, Bekker et al.19 reported modulation of inflammatory mediators and improved self-reported quality of recovery after major spine surgery with intraoperative dexmedetomidine. Analgesic consumption and pain rating were not reported in their study but the possibility of a pleotropic effect of dexmedetomidine cannot be excluded. We did not measure quality of recovery or patient satisfaction in this study.
One of the major side effects of chronic opioid therapy is the development of opioid tolerance and/or opioid-induced hyperalgesia (OIH).20 Opioid tolerance is characterized by a decreased response to the opioid’s analgesic effects over time with ultimate loss of analgesic efficacy, whereas in OIH, an enhanced pain response to noxious stimuli is evident after both acute and chronic opioid administration.21 OIH may develop even after short periods of intraoperative opioid administration reducing the efficacy of postoperative opioids.22,23 This risk of opioid tolerance and OIH is particularly high in patients with back pain receiving chronic opioid therapy.24,25
Synergistic use of an antihyperalgesic agent can reduce the incidence of opioid tolerance and OIH potentially reducing the amount of postoperative opioids. Dexmedetomidine modulates hyperalgesia by increasing the spinal acetylcholine levels and by modulating the spinal cord NMDA receptor subtype 2B.26,27 The effects of dexmedetomidine on OIH in major spine surgery have not been adequately studied. In contrast other non-opioid drugs, such as ketamine and lidocaine, that modulate OIH have demonstrated opioid-sparing effects in spine surgery.
Loftus et al.18 found reduced opioid consumption for up to 6 weeks after multilevel spine surgery in opioid-dependent chronic pain patients who received intraoperative ketamine. These beneficial effects were accrued without an increased incidence of side effects.
In contrast to ketamine, lidocaine is thought to modify OIH by inhibition of conventional protein kinase C gamma.28 Lidocaine administered intraoperatively and continued for up to 8 hours postoperatively reduces opioid use with better pain ratings after spine surgery. Interestingly the beneficial effects of lidocaine are evident up to 3 months after surgery when evaluated by the short form-12 physical composite score.2 This sustained functional improvement after an intraoperative intervention that modulates inflammation and hyperalgesia is similar to the results reported in the study by Bekker et al. mentioned previously.19 Of note in the aforementioned lidocaine study, 66% were instrumented cases with median 4 levels of fusion suggesting significant tissue injury.2
One of the weaknesses of this study was that it was terminated early after interim analysis crossed a futility boundary. To demonstrate the effect size originally hypothesized (30% reduction in median opioid consumption after surgery), continuing the study to full recruitment would have only a 9% power. Early termination of studies can happen when the hypothesized effect size is not evident or the accrual rate is slow.29 In our study, both situations were encountered. Our recruitment period to interim analysis spanned 3 years, and planned practice changes to standardize perioperative care for major spine surgery made it unfeasible to continue the study.
An additional weakness was postoperative analgesia, which was not strictly standardized and was at the discretion of the intensive care team. Nevertheless, postoperative pain control was initiated using a standard admission pain order set. Although the ICU staff were blinded to the treatment group, the risk of variable treatment thresholds for clinician-driven pain medication administration was possible in this study. However, all patients were managed in the same ICU in an attempt to standardize care and rapidly transitioned to patient-controlled analgesia.
Finally, a fixed dose of dexmedetomidine was used with no stratification for a low-dose versus high-dose group. Using a single-dose study design can miss the possibility of a dose-responsive effect.
Although previous studies have reported opioid sparing with dexmedetomidine, this benefit is not evident with adult deformity correction surgery. This prospective randomized double-blind study shows that intraoperative dexmedetomidine does not reduce opioid consumption or improve pain scores after multilevel spine surgery. There is a slight reduction in intraoperative opioid requirements, but this effect is not sustained in the postoperative period.
Name: Bhiken I. Naik, MBBCh.
Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.
Attestation: Bhiken I. Naik has seen the original study data, reviewed the analysis of the data, approved the final manuscript, and is the author responsible for archiving the study files.
Name: Edward C. Nemergut, MD.
Contribution: This author helped design the study, analyze the data, and write the manuscript.
Attestation: Edward C. Nemergut has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.
Name: Ali Kazemi, MD.
Contribution: This author helped conduct the study.
Attestation: Ali Kazemi has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.
Name: Lucas Fernández, MD, DSc.
Contribution: This author helped conduct the study.
Attestation: Lucas Fernández has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.
Name: Sarah K. Cederholm, MD.
Contribution: This author helped conduct the study.
Attestation: Sarah K. Cederholm has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.
Name: Timothy L. McMurry, PhD.
Contribution: This author helped analyze the data.
Attestation: Timothy L. McMurry has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.
Name: Marcel E. Durieux, MD, PhD.
Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.
Attestation: Marcel E. Durieux has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.
Dr. Edward C. Nemergut is the Section Editor for Graduate Medical Education for Anesthesia & Analgesia. This manuscript was handled by Spencer S. Liu, MD, and Dr. Nemergut was not involved in any way with the editorial process or decision.
1. Fu KM, Rhagavan P, Shaffrey CI, Chernavvsky DR, Smith JS. Prevalence, severity, and impact of foraminal and canal stenosis among adults with degenerative scoliosis. Neurosurgery. 2011;69:1181–7
2. Farag E, Ghobrial M, Sessler DI, Dalton JE, Liu J, Lee JH, Zaky S, Benzel E, Bingaman W, Kurz A. Effect of perioperative intravenous lidocaine administration on pain, opioid consumption, and quality of life after complex spine surgery. Anesthesiology. 2013;119:932–40
3. Sharma S, Balireddy RK, Vorenkamp KE, Durieux ME. Beyond opioid patient-controlled analgesia: a systematic review of analgesia after major spine surgery. Reg Anesth Pain Med. 2012;37:79–98
4. Grewal A. Dexmedetomidine: new avenues. J Anaesthesiol Clin Pharmacol. 2011;27:297–302
5. Gunes Y, Tevfik Hayri O, Gunduz HM, Gedik YE, Erman T, Ozcengiz D. Patient-controlled analgesia comparison of morphine to dexmedetomidine plus morphine in patients undergoing laminectomy. Neurosurg Q. 2008;18:178–81
6. Ozkose Z, Demir FS, Pampal K, Yardim S. Hemodynamic and anesthetic advantages of dexmedetomidine, an alpha 2-agonist, for surgery in prone position. Tohoku J Exp Med. 2006;210:153–60
7. Koinig H, Wallner T, Marhofer P, Andel H, Horauf K, Mayer N. Magnesium sulfate reduces intra- and postoperative analgesic requirements. Anesth Analg. 1998;87:206–10
8. Mirza SK, Deyo RA, Heagerty PJ, Turner JA, Lee LA, Goodkin R. Towards standardized measurement of adverse events in spine surgery: conceptual model and pilot evaluation. BMC Musculoskelet Disord. 2006;7:53
9. Pocock S. Group sequential methods in the design and analysis of clinical trials. Biometrika. 1977;64:191–9
10. Dexter F. Wilcoxon-Mann-Whitney test used for data that are not normally distributed. Anesth Analg. 2013;117:537–8
11. Divine G, Norton HJ, Hunt R, Dienemann J. Statistical grand rounds: a review of analysis and sample size calculation considerations for Wilcoxon tests. Anesth Analg. 2013;117:699–710
12. Bakhamees HS, El-Halafawy YM, El-Kerdawy HM, Gouda NM, Altemyatt S. Effects of dexmedetomidine in morbidly obese patients undergoing laparoscopic gastric bypass. Middle East J Anaesthesiol. 2007;19:537–51
13. Tufanogullari B, White PF, Peixoto MP, Kianpour D, Lacour T, Griffin J, Skrivanek G, Macaluso A, Shah M, Provost DA. Dexmedetomidine infusion during laparoscopic bariatric surgery: the effect on recovery outcome variables. Anesth Analg. 2008;106:1741–8
14. Turgut N, Turkmen A, Gökkaya S, Altan A, Hatiboglu MA. Dexmedetomidine-based versus fentanyl-based total intravenous anesthesia for lumbar laminectomy. Minerva Anestesiol. 2008;74:469–74
15. Yildiz M, Tavlan A, Tuncer S, Reisli R, Yosunkaya A, Otelcioglu S. Effect of dexmedetomidine on haemodynamic responses to laryngoscopy and intubation: perioperative haemodynamics and anaesthetic requirements. Drugs R D. 2006;7:43–52
16. Blaudszun G, Lysakowski C, Elia N, Tramèr MR. Effect of perioperative systemic α2 agonists on postoperative morphine consumption and pain intensity: systematic review and meta-analysis of randomized controlled trials. Anesthesiology. 2012;116:1312–22
17. Gottschalk A, Durieux ME, Nemergut EC. Intraoperative methadone improves postoperative pain control in patients undergoing complex spine surgery. Anesth Analg. 2011;112:218–23
18. Loftus RW, Yeager MP, Clark JA, Brown JR, Abdu WA, Sengupta DK, Beach ML. Intraoperative ketamine reduces perioperative opiate consumption in opiate-dependent patients with chronic back pain undergoing back surgery. Anesthesiology. 2010;113:639–46
19. Bekker A, Haile M, Kline R, Didehvar S, Babu R, Martiniuk F, Urban M. The effect of intraoperative infusion of dexmedetomidine on the quality of recovery after major spinal surgery. J Neurosurg Anesthesiol. 2013;25:16–24
20. Lee M, Silverman SM, Hansen H, Patel VB, Manchikanti L. A comprehensive review of opioid-induced hyperalgesia. Pain Physician. 2011;14:145–61
21. Lee HJ, Yeomans DC. Opioid induced hyperalgesia in anesthetic settings. Korean J Anesthesiol. 2014;67:299–304
22. Guignard B, Bossard AE, Coste C, Sessler DI, Lebrault C, Alfonsi P, Fletcher D, Chauvin M. Acute opioid tolerance: intraoperative remifentanil increases postoperative pain and morphine requirement. Anesthesiology. 2000;93:409–17
23. Luginbühl M, Gerber A, Schnider TW, Petersen-Felix S, Arendt-Nielsen L, Curatolo M. Modulation of remifentanil-induced analgesia, hyperalgesia, and tolerance by small-dose ketamine in humans. Anesth Analg. 2003;96:726–32
24. Chaparro LE, Furlan AD, Deshpande A, Mailis-Gagnon A, Atlas S, Turk DC. Opioids compared to placebo or other treatments for chronic low-back pain. Cochrane Database Syst Rev. 2013;8:CD004959
25. Chu LF, D’Arcy N, Brady C, Zamora AK, Young CA, Kim JE, Clemenson AM, Angst MS, Clark JD. Analgesic tolerance without demonstrable opioid-induced hyperalgesia: a double-blinded, randomized, placebo-controlled trial of sustained-release morphine for treatment of chronic nonradicular low-back pain. Pain. 2012;153:1583–92
26. Kimura M, Saito S, Obata H. Dexmedetomidine decreases hyperalgesia in neuropathic pain by increasing acetylcholine in the spinal cord. Neurosci Lett. 2012;529:70–4
27. Zheng Y, Cui S, Liu Y, Zhang J, Zhang W, Zhang J, Gu X, Ma Z. Dexmedetomidine prevents remifentanil-induced postoperative hyperalgesia and decreases spinal tyrosine phosphorylation of N-methyl-d-aspartate receptor 2B subunit. Brain Res Bull. 2012;87:427–31
28. Cui W, Li Y, Li S, Yang W, Jiang J, Han S, Li J. Systemic lidocaine inhibits remifentanil-induced hyperalgesia via the inhibition of cPKCgamma membrane translocation in spinal dorsal horn of rats. J Neurosurg Anesthesiol. 2009;21:318–25
29. van der Tweel I, van Noord PA. Early stopping in clinical trials and epidemiologic studies for “futility”: conditional power versus sequential analysis. J Clin Epidemiol. 2003;56:610–7