Opioid-Free Anesthesia for Craniotomy : Journal of Neurosurgical Anesthesiology

Secondary Logo

Journal Logo

Short Reports

Opioid-Free Anesthesia for Craniotomy

McCullough, Ian L. MD*; Shteamer, Jack W. MD*; Erwood, Andrew M. MD; Spektor, Boris MD*; Boorman, David W. MS*; Sharifpour, Milad MD*; Olson, Jeffery J. MD; Papangelou, Alexander MD*

Author Information
Journal of Neurosurgical Anesthesiology 35(1):p 80-85, January 2023. | DOI: 10.1097/ANA.0000000000000797


Opioids are considered the mainstay of perioperative pain management; however, the adverse effects of opioids are particularly problematic following craniotomy as they can lead to an inaccurate neurological examination because of excessive sedation and have the potential to mask early signs of intracranial complications. Similarly, opioid-induced respiratory depression promotes hypercapnia which increases cerebral blood flow and may lead to cerebral edema.1 Conversely, suboptimal analgesia drives sympathetic efflux, promoting hypertension that may increase morbidity and end organ damage through intracranial hemorrhage. Despite traditional opioid-based analgesia, moderate to severe postoperative pain occurs in over two-thirds of craniotomy patients, and there is a strong impetus to optimize postcraniotomy pain management with nonopioid analgesic agents.1,2

International clinical practice guidelines for the management of postoperative pain support the use of multimodal-analgesia, defined as the use of analgesics with different mechanisms of action in combination with potential peripheral or central nerve blockade to achieve additive or synergistic perioperative pain relief.3 A meta-analysis in 2013 demonstrated benefits of preincisional scalp block in reducing pain scores up to 8 hours and morphine consumption up to 24 hours after craniotomy, supporting its use for perioperative analgesia in supratentorial craniotomy.4 Intravenous (IV) acetaminophen has also shown modest benefits, most importantly in reducing the incidence of severe pain following craniotomy (50% vs. 26%) and increasing patient satisfaction.5 Dexmedetomidine infusion has also shown compelling analgesic activity during craniotomy.6 In addition to demonstrating lower pain scores and reduced opioid consumption for up to 12 hours after surgery, there were fewer opioid-related side effects and no difference in sedation.7,8 Given the availability and effectiveness of these nonopioid analgesic agents as well as the growing interest in opioid-free anesthesia, we hypothesized that a practical, multimodal, opioid-free anesthetic (OFA) regimen comprising preoperative scalp block combined with IV dexmedetomidine infusion and IV acetaminophen would be feasible and noninferior to a more conventional opioid-based anesthetic technique.


This study was approved by the Emory University Institutional Review Board (STUDY00000960). We conducted a mixed prospective-retrospective cohort study at Emory University Hospital (EUH), comparing patients who were prospectively identified for opioid-free craniotomies to control patients who were identified retrospectively after undergoing craniotomy utilizing opioid-based anesthetics. Predefined primary outcomes include morphine milligram equivalents (MME) consumption and pain scores (numerical rating scale, 0 to 10 points) assessed at intervals of time: in postanesthesia care unit (PACU), and 0 to 12 hours, 12 to 24 hours, and 0 to 24 hours postoperatively. Our secondary outcomes include time from leaving the operating room to PACU discharge (Recovery), time to first rescue opioid, as well as end of surgery to extubation (Emergence) time. We have also reported on hemodynamic response to cranial pinning and surgical incision.

In the OFA technique, all patients underwent anatomically directed scalp block of the supraorbital, supratrochlear, auriculotemporal, posterior branch of the great auricular, greater occipital, and lesser occipital nerves.9,10 The zygomaticotemporal nerve was also blocked if indicated. Five patients received scalp blocks with 0.5% bupivacaine, and the sixth patient received a mixture of 0.5% bupivacaine and Exparel (bupivacaine liposome). The scalp block was placed before incision but after cranial pinning to facilitate neuronavigation. Dexmedetomidine infusion was initiated upon entering the operating room at 0.5 mcg/kg/h and discontinued at skin closure.7,8 Induction of anesthesia was performed with a combination of propofol, esmolol, and a neuromuscular blocking agent. Additional propofol or esmolol were administered before cranial pinning to blunt the hemodynamic response.11 Three patients in the OFA group received only sevoflurane for maintenance anesthesia, while 3 patients received a combination of 1/2 minimum alveolar concentration of sevoflurane combined with propofol infusion to facilitate neuromonitoring. IV acetaminophen 1000 mg was administered before surgical closure.12

The OFA arm of the study included 3 males and 3 females between the ages of 30 and 80 years who underwent elective supratentorial craniotomy for tumor resection at EUH between August 2019 and June 2020. Postoperative magnetic resonance images were analyzed to determine incision length and location, facilitating identification of 4 incision categories: frontotemporoparietal (FTP), frontotemporal (FT), temporoparietal (TP), and parietal (P). Only P incisions were free from temporalis muscle involvement. All of the OFA patients were identified prospectively. They were first-start supratentorial craniotomies for tumor resection, allowing for completion by a single anesthesiologist. Other than avoidance of vascular neurosurgical cases, no exclusion criteria were used.

Eighteen controls were selected from a database containing all 943 craniotomies performed at EUH from 2013 to 2015. Patients were randomized and simultaneously matched for equal ratios of incision locations (FTP, FT, TP, and P) and sex, followed by matching by age and incision length (Fig. 1). We did not place any exclusion criteria on the analgesic approach used by the anesthesia team; however, all control patients received fentanyl as their only intraoperative opioid. The 3:1 ratio of controls to OFA patients was chosen to afford greater power by providing a more reliable average in the control group for comparison with the OFA group.

Demonstrating the selection process of the 18 control patients matched by sex, incision location, incision length and age.

Statistical Analysis and Sample Size Calculation

Data were analyzed using SAS 9.4 (Cary, NC). The Mann-Whitney U test and the Fisher exact test were used with patient demographics and confounders to confirm the 2 groups were not statistically different. Comparison of variances was performed using the F test of equal variances. Noninferiority testing was used to determine if the OFA group was noninferior in outcome measures.13 In this method, a noninferiority margin (Delta) is declared in advance to be a “clinically relevant” difference and added to the control mean. If the 95% upper confidence limit of the OFA group is lower than the control mean plus delta, the groups are deemed sufficiently similar. A distribution-free 95% confidence interval (CI) was calculated, to avoid assumptions of normality.

For postoperative opioid consumption, a delta of 25 MME was selected for each time interval of interest.14 Delta was set at 10 MME in PACU because of the shorter duration. Delta for postoperative pain scores was set at 2 based on a systematic review showing the minimum clinically important difference in acute pain to have an interquartile range of 1.4 to 2.3.15 Delta for recovery time and time to first rescue opioid were each set at 1 hour. Delta for emergence was set at 10 minutes (0.167 h). A sensitivity statistical power analysis was conducted.16 With a noninferiority margin for MME set at 25 and an experimental average SD of 17, a sample size of 6 per group would be sufficient to determine noninferiority with 80% certainty.

Mean arterial pressure (MAP) and heart rate (HR) were extracted from medical records and recorded immediately prior to and following both cranial pinning and incision. We set a MAP of 130 mmHg and HR of 110 beats/min as critical thresholds above which were managed with propofol or esmolol boluses.


The 2 groups were successfully matched by age and incision location, and were statistically similar by body mass index, length of stay, number of pain scores recorded, time on anesthesia, history of chronic pain, home opioid use, and propofol infusion (Table 1).

TABLE 1 - Surgical and Patient Variables of Interest
Mean (SD)
Continuous Variable Opioid-free Control P *
Incision length (cm) 21.2 (6.0) 21.9 (9.0) 0.97
Age (y) 51.3 (14.5) 52.7 (16.6) 0.84
BMI (m/kg2) 30.2 (10.9) 30.2 (6.7) 0.52
Length of stay (h)
 Surgery start-stop 2.12 (0.70) 2.52 (1.40) 0.49
 Time in-out 3.68 (0.73) 3.83 (1.61) 0.67
Number of pain scores recorded
 0-12 h 9.17 (4.92) 9.78 (3.12) 0.51
 12- 24 h 8.00 (1.67) 6.83 (1.76) 0.24
Time on anesthesia (h) 3.20 (0.86) 3.48 (1.61) 0.92
Categorical Variable Group N (%) N (%)
Sample size Total 6 (100) 18 (100) NA
Incision location§ FTP 2 (33) 6 (33) Matched
FT 2 (33) 6 (33)
P # 1 (17) 3 (17)
TP ** 1 (17) 3 (17)
Sex Female 3 (50) 9 (50) Matched
Chronic pain Yes 1 (17) 1 (5.6) 0.45
Home opioid use Yes 1 (17) 1 (5.6) 0.45
Propofol infusion Yes 3 (50) 2 (11) 0.079
The control group was successfully matched by incision length and age, and exactly matched by incision location and sex. The 2 groups were also similar by chronic pain history, BMI, length of stay and home opioid use.
*Mann-Whitney U test, nonparametric equivalent to the t test.
Fisher exact test.
§F: Frontal, T: Temporal, P: Parietal.
FTP included patients with a “question mark” or “reverse question mark” incision.
FT included patients with a curvilinear or “pterional” incision.
#TP included patients with an inverted “U” or “horseshoe” shaped incision.
**P included patients with a “linear” or “curvilinear” incision.
BMI indicates body mass index; FT, frontotemporal; FTP, frontotemporoparietal; NA, not applicable; P, parietal; TP, temporoparietal.

Noninferiority of the OFA technique was demonstrated for MME consumption at all postoperative time intervals of interest (Table 2) (Supplemental Figure 1, Supplemental Digital Content, https://links.lww.com/JNA/A419). Noninferiority was shown with respect to postoperative pain scores in the PACU, from 0 to 12 hours, from 0 to 24 hours, but not for the 12-hour to 24-hour period (Fig. 2) Supplemental Figure 2, Supplemet Digital Content 2, https://links.lww.com/JNA/A420). Noninferiority was also established for recovery time but not emergence time Supplemental Figures 3 and 4, Supplemental Digital Content 3 and 4, https://links.lww.com/JNA/A421; https://links.lww.com/JNA/A422).

TABLE 2 - Comparison of Opioid-free Group to Controls and Noninferiority Margin (Δ)
Opioid-free Noninferiority Control Control
Variable Time Mean 95% CI Shown* Δ Mean±Δ Mean 95% CI
Average MME (mg) PACU , 3.50 0.0-14.0 Yes 10 17.83 7.83 0.0-10.0
0-12 h 23.1 7.5-35.0 Yes 25 55.9 30.9 12.0-46.0
12-24 h 23.0 12.0-39.0 Yes 25 48.8 23.8 6.0-36.0
0-24 h 23.0 11.3-33.8 Yes 25 52.3 27.3 9.0-38.0
Average pain score PACU , 0.28 0.00-1.12 Yes 2 6.53 4.53 0.75-6.80
0-12 h 2.57 0.75-4.60 Yes 2 5.36 3.36 2.15-4.46
12-24 h 3.11 0.80-6.00 No 2 4.04 2.04 0.83-3.67
0-24 h 2.84 1.05-4.60 Yes 2 4.67 2.67 1.49-3.81
Emergence (h)
 End of surgery to extubation 0.24 0.08-0.47 No 0.17 0.33 0.16 0.08-0.20
Recovery (h)
 Leaving OR to leaving PACU 1.98 1.05-3.22 Yes 1.0 3.41 2.41 1.30-2.57
Time to first rescue opioid (h) § 4.98 0.37-8.93 Border 1.0 0.71 1.71 0.63-2.38
To identify noninferiority, compare the upper confidence limit of the opioid-free group to the mean of the control group+delta. Examine the lower confidence limit of the when a low number is undesirable, that is, the time to first rescue opioid.
*Is noninferiority demonstrated? That is, is the upper control limit of the opioid-free group less than control mean+delta?
Excludes N=2 in opioid-free group with no PACU time.
Data overlap with data for 0 to 12 hours and 0 to 24 hours.
§For this variable, longer is better, so examine if the lower control limit of the opioid-free group greater than the mean–delta.
Not shown under a distribution-free interval (0.37 to 8.93) but shown under a normal confidence interval (1.53-8.42).
CI indicates confidence interval; MME, morphine milligram equivalents; OR, operating room; PACU, postanesthesia care unit; Δ, noninferiority margin.

Boxplots overlaid with individual data points for the opioid-free and control groups at measured time points for both morphine milligram equivalent (A) and average pain scores (B). B, Noninferiority is demonstrated for morphine milligram equivalent at all-time points, and for average pain score at all-time points except 12 to 24 hours. Two patients in the opioid-free group were sent directly to the intensive care unit and bypassed the PACU. Rectangles reflect the Q1, median and Q3 of the data, the diamond reflects the mean, while data points beyond the whiskers reflect outliers. PACU indicates postanesthesia care unit.

Noninferiority was inconclusive for time to first rescue opioid time Supplemental Figure 3, Supplemental Digital Content 3, https://links.lww.com/JNA/A421); it was not shown using a distribution-free CI (0.38-8.93 vs. 0.71) but was shown with a normal CI (1.53-8.42 vs. 0.71). Failure using the distribution-free CI was driven by a single patient. The variance of the OFA group was also significantly greater than the control group (P=0.017).

Immediately following cranial pinning the proportion of subjects exceeding the HR threshold of 110 beats/min was 1/6 for the OFA group and 2/18 for the control group (Supplemental Figure 5, Supplement Digital Content 5, https://links.lww.com/JNA/A423). The proportion exceeding the MAP threshold of 130 mm Hg following cranial pinning was 1/6 for the OFA group and 3/18 for the control group. No patients in either group experienced HR >110 beats/min or MAPs >130 mm Hg following surgical incision (Supplemental Figure 6, Supplement Digital Content, https://links.lww.com/JNA/A424).


This study confirms the feasibility of an OFA technique in patients undergoing supratentorial craniotomy for tumor resection. We demonstrated noninferiority with respect to postoperative opioid consumption at all measured time points: in the PACU, 0 to 12 hours, 12 to 24 hours, and 0 to 24 hours postoperatively. Noninferiority was further demonstrated for postoperative pain scores 0 to 12 hours and 0 to 24 hours, but not for 12 to 24 hours postoperatively. We were also able to show noninferiority for recovery time, but time to administration of the first rescue opioid was borderline depending on the type of CI used. Finally, noninferiority was not achieved for emergence time with the OFA group taking on average 4.2 minutes longer to emerge relative to the control group. It is possible this increase in emergence time resulted from the use of dexmedetomidine. However, prolonged emergence time was not seen by Peng et al,8 whose dexmedetomidine dosing regimen we used.

One concern in utilizing scalp block as the cornerstone of the OFA technique is the duration of postoperative analgesia. A recent meta-analysis concluded efficacy for a preincisional block may only last up to 8 hours postoperatively.4 In this study, the patients who received OFA had a small but notable increase in their pain scores at 12 to 24 hours postoperatively compared with controls. This may be attributed to the diminishing analgesic effect of scalp blockade. Further studies incorporating liposomal bupivacaine into the preincisional block would be of interest, as would the integration of one or more agents with distinct and complementary analgesic properties such as ketamine, magnesium sulfate, gabapentinoids or selective Cox-2 inhibitors into the OFA technique.17,18

When considering the physiological response to cranial pinning and incision, concern should rest with extreme hemodynamic responses that may lead to patient harm. To our knowledge, an absolute threshold for risk does not exist, and, clinically, an individualized threshold should supersede any recommendation. Our results do not reveal any notable difference between the OFA and control groups with respect to our clinically significant hemodynamic thresholds. In the setting of a prospective randomized trial, we would plan to create a protocol to include scalp blockade before pinning and more standardized anesthetic dosing before to cranial pinning and surgical incision. Evidence exists supporting scalp blockade before head fixation as a means of attenuating the hemodynamic response.19,20

Limitations of this study include its small sample size and retrospective component. Moreover, the control patients underwent general anesthesia several years before those who received OFA, and the difference in postoperative MME consumption could reflect not only the change in anesthetic technique but also changes in opioid administration in light of the ongoing opioid epidemic. Another possible limitation is that all control patients received only fentanyl for intraoperative analgesia, which may not be representative of current neuroanesthesiology practice. Lastly, given the structure of our study, it was impossible to standardize postoperative pain assessment and opioid administration in the PACU.

This study demonstrates the feasibility of the OFA technique for patients undergoing supratentorial craniotomy and should serve as pilot data for further research. Future multicenter prospective randomized clinical trials could investigate differences in outcomes such as time to emergence from general anesthesia, postoperative sedation score, effect on postoperative neurological examination, incidence of postoperative respiratory depression, postoperative hemodynamics, incidence of postoperative nausea and vomiting, the quality of analgesia, postoperative opioid consumption, and overall patient satisfaction. Intraoperative processed electroencephalography may be of interest to obtain data on the depth of anesthesia and electroencephalographic response to pinning and incision. Studies could also evaluate potential differences in chronic incisional pain and chronic opioid use as a result of analgesic technique. Similarly, it is imperative to develop multidisciplinary, evidence-based protocols for assessment and management of postcraniotomy pain in the PACU and intensive care unit, with the goal of decreasing opioid utilization to minimize their associated undesirable side effects.


The authors would like to acknowledge Allan Gottschalk for his assistance with manuscript editing.


1. Dunn LK, Naik BI, Nemergut EC, et al. Post-craniotomy pain management: beyond opioids. Curr Neurol Neurosci Rep. 2016;16:93. doi:10.1007/s11910-016-0693-y
2. Gottschalk A, Berkow LC, Stevens RD, et al. Prospective evaluation of pain and analgesic use following major elective intracranial surgery. J Neurosurg. 2007;106:210–216. doi:10.3171/jns.2007.106.2.210
3. Chou R, Gordon DB, de Leon-Casasola OA, et al. Management of postoperative pain: a clinical practice guideline from the American Pain Society, the American Society of Regional Anesthesia and Pain Medicine, and the American Society of Anesthesiologists’ Committee on Regional Anesthesia, Executive Committee, and Administrative Council. J Pain. 2016;17:131–157. doi:10.1016/j.jpain.2015.12.008
4. Guilfoyle MR, Helmy A, Duane D, et al. Regional scalp block for postcraniotomy analgesia: a systematic review and meta-analysis. Anesth Analg. 2013;116:1093–1102. doi:10.1213/ANE.0b013e3182863c22
5. Artime CA, Aijazi H, Zhang H, et al. Scheduled intravenous acetaminophen improves patient satisfaction with postcraniotomy pain management: a prospective, randomized, placebo-controlled, double-blind study. J Neurosurg Anesthesiol. 2018;30:231–236. doi:10.1097/ANA.0000000000000461
6. Rajan S, Hutcherson MT, Sessler DI, et al. The effects of dexmedetomidine and remifentanil on hemodynamic stability and analgesic requirement after craniotomy: a randomized controlled trial. J Neurosurg Anesthesiol. 2016;28:282–290. doi:10.1097/ANA.0000000000000221
7. Song J, Ji Q, Sun Q, et al. The opioid-sparing effect of intraoperative dexmedetomidine infusion after craniotomy. J Neurosurg Anesthesiol. 2016;28:14–20. doi:10.1097/ANA.0000000000000190
8. Peng K, Jin X-h, Liu SI, et al. Effect of intraoperative dexmedetomidine on post-craniotomy pain. Clin Ther. 2015;37:1114–1121.e1. doi:10.1016/j.clinthera.2015.02.011
9. Nguyen A, Girard F, Boudreault D, et al. Scalp nerve blocks decrease the severity of pain after craniotomy. Anesth Analg. 2001;93:1272–1276. doi:10.1097/00000539-200111000-00048
10. Osborn I, Sebeo J. “Scalp block” during craniotomy: a classic technique revisited. J Neurosurg Anesthesiol. 2010;22:187–194. doi:10.1097/ANA.0b013e3181d48846
11. Wiest DB, Haney JS. Clinical pharmacokinetics and therapeutic efficacy of esmolol. Clin Pharmacokinet. 2012;51:347–356. doi:10.2165/11631590-000000000-00000
12. Sivakumar W, Jensen M, Martinez J, et al. Intravenous acetaminophen for postoperative supratentorial craniotomy pain: a prospective, randomized, double-blinded, placebo-controlled trial. J Neurosurg. 2018;130:766–772. doi:10.3171/2017.10.JNS171464
13. Walker E, Nowacki AS. Understanding equivalence and noninferiority testing. J Gen Intern Med. 2011;26:192–196. doi:10.1007/s11606-010-1513-8
14. Hussain N, Costache I, Kumar N, et al. Is supraclavicular block as good as interscalene block for acute pain control following shoulder surgery? A systematic review and meta-analysis. Anesth Analg. 2020;130:1304–1319. doi:10.1213/ANE.0000000000004692
15. Olsen MF, Bjerre E, Hansen MD, et al. Pain relief that matters to patients: systematic review of empirical studies assessing the minimum clinically important difference in acute pain. BMC Med. 2017;15:35. doi:10.1186/s12916-016-0775-3
16. Faul F, Erdfelder E, Lang A-G, et al. G*Power 3: a flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behav Res Methods. 2007;39:175–191. doi:10.3758/bf03193146
17. Mahajan C, Mishra RK, Jena BR, et al. Effect of magnesium and lignocaine on post-craniotomy pain: a comparative, randomized, double blind, placebo-controlled study. Saudi J Anaesth. 2019;13:299–305. doi:10.4103/sja.SJA_837_18
18. Ban VS, Bhoja R, McDonagh DL. Multimodal analgesia for craniotomy. Curr Opin Anaesthesiol. 2019;32:592–599. doi:10.1097/ACO.0000000000000766
19. Pinosky ML, Fishman RL, Reeves ST, et al. The effect of bupivacaine skull block on the hemodynamic response to craniotomy. Anesth Analg. 1996;83:1256–1261. doi:10.1213/00000539-199612000-00022
20. Geze S, Yilmaz AA, Tuzuner F. The effect of scalp block and local infiltration on the haemodynamic and stress response to skull-pin placement for craniotomy. Eur J Anaesthesiol. 2009;26:298–303. doi:10.1097/EJA.0b013e32831aedb2

multimodal-analgesia; opioid-free-craniotomy; postoperative pain; scalp block

Supplemental Digital Content

Copyright © 2021 Wolters Kluwer Health, Inc. All rights reserved.