Postoperative Analgesia in Lumbar Discectomy: Morphine vs. Dexamethasone-Soaked Epidural Gelfoam: A Randomized Controlled Trial

Abstract

Background

This study compared the analgesic efficacy of gelfoam-soaked epidural morphine with gelfoam-soaked epidural dexamethasone following lumbar discectomies as their clinical comparison is lacking.

Materials and Methods

Sixty patients undergoing single-level lumbar discectomy surgeries were randomized either in group M (gelfoam-soaked epidural morphine, 3 mg) or in group D (gelfoam-soaked epidural dexamethasone, 8 mg) placed over the duramater before the closure. The primary objective was to compare the mean time to first rescue analgesia for 24 postoperative hours. The secondary objectives were pain intensity, total rescue analgesic requirement, and to look for any adverse event related to study drugs.

Statistical Analysis

The primary outcome was analyzed using a two-tailed unpaired t -test as well as a Cox proportional hazards model. The categorical outcome of complete response was compared using a chi-squared test. The other numerical outcomes were compared using a linear mixed effects model accounting for repeated measures and the categorical repeated measure outcome was evaluated using a binomial generalized estimating equations model.

Results

The mean time to first analgesic (hours) in both the groups (M vs. D, 26 [12] vs. 19 [6], mean difference = 7 [95% confidence interval [CI], −8.2, 22, p  = 0.3, unpaired t -test) was not significant. The pain scores showed a tendency toward low pain scores at 24 hours in the morphine group. The sedation scores and number needing rescue analgesia did not differ between groups. Pain scores and the number needing rescue analgesia decreased significantly in both groups over time. There was a difference in terms of complete response between the groups (M vs. D, 83% vs. 57%, difference of 27% [95% CI of 1.1%, 52%], p  = 0.02, chi-square test).

Conclusion

Gelfoam-soaked epidural morphine versus dexamethasone demonstrated no difference in mean time to first rescue analgesic following lumbar discectomy surgeries but gelfoam-soaked epidural morphine reduced the requirement of rescue analgesia in the first 24 postoperative hours compared with gelfoam-soaked epidural dexamethasone.

Introduction

Lumbar discectomy is associated with severe pain due to the shearing and traction of muscles, bony dissection, and may have postoperative sequalae if adequate analgesia is not provided in perioperative period. A variety of analgesics have been tried to control this pain, not only by the parenteral route but also through the central neuraxial route. Intravenous opioids are considered as one of the best modalities of postoperative analgesia following lumbar discectomy. ^{1 2} But the problems associated with parenteral opioid administration are the systemic effects and the associated adverse effects.

Epidural opioids, both fentanyl and morphine, have been tried in spine surgeries providing adequate pain control and fewer side effects compared with the parenteral route. ^{3 4} The advantages of epidural fentanyl are the early onset, greater potency, less sedation, and respiratory depression. However, its analgesic effect is short lived making it unsuitable for prolonged postoperative analgesia. ⁵ Single dose of opioids infiltrated over the exposed duramater prior to closure have the disadvantage of a short duration of action. This can be addressed by use of gelfoam soaked with opioids in the epidural space. ^{6 7 8} Epidural morphine has been given as single dose due to its long duration of action providing prolonged analgesia following spine surgery. Morphine has also been administered through epidural catheters, but this practice is associated with the risk of local site infection and is also difficult to manage following spine surgery. ^{9 10 11}

Beside epidural morphine and dexamethasone injections have also been used to control postoperative pain control following lumbar spine surgery. The pain control with epidural dexamethasone has been claimed to be good and devoid of associated side effects, ^{12 13} but limited literature has compared epidural morphine against dexamethasone to control postoperative pain after spine surgeries.

This study was conducted in patients undergoing single-level lumbar spine surgery to compare the difference in analgesic requirements in the first 24 postoperative hours following intraoperative epidural implantation of gelfoam-soaked morphine against gelfoam-soaked dexamethasone. The study hypothesized that the analgesic effect of epidural gelfoam-soaked morphine is better as compare with epidural gelfoam-soaked dexamethasone following lumber discectomy surgery.

The primary objective was to compare the time to first rescue analgesia. The secondary objectives were pain intensity, total rescue analgesic requirement, and to look for any adverse event related to study drugs or surgery.

Materials and Methods

This prospective, double-blinded, randomized control study was conducted in the department of anesthesiology and critical care at a tertiary care center from March 2020 to August 2024 following institutional ethical approval (IEC/2018/601) and registration in the Clinical Trials Registry - India (CTRI/2020/02/023316; dated 13–02–2020). A written informed consent was obtained from all the study participants. The article adheres to the applicable CONSORT guidelines. Study participants included patients of either gender with age 18 to 65 years undergoing single-level lumbar discectomy surgeries. Exclusion criteria were refusal to participate in the study, body mass index below 18 or above 25, cervical or thoracic level laminectomies, spine fixation surgeries, history of previous spine surgery, patients with neuromuscular and psychological diseases, accidental dural breach, patients on systemic steroids, or allergic to study drugs.

All the participants underwent preanesthetic evaluation before surgery and the visual analog scale (VAS) score for pain was explained. In the operating room, standard monitors (electroencephalogram, noninvasive blood pressure, pulse oximeter) were attached. The participants were premedicated with injection midazolam 1 mg after securing the intravenous cannula in the nondominant upper limb. The participants were induced with injection propofol (1–2 mg/kg) till loss of response to verbal command and the intraoperative analgesia was maintained with injection fentanyl 2 µg/kg bolus at induction along with 1 µg/kg/hour infusion during maintenance anesthesia. The muscle relaxation was achieved with rocuronium (0.8 mg/kg) bolus following induction and maintained targeting a zero to one train of four count. The trachea was intubated using an armored endotracheal tube of appropriate size and controlled mechanical ventilation with volume control mode (tidal volume 6 mL/kg; respiratory rate 12/min; inspiratory:expiratory ratio 1:2; positive end expiratory pressure – 5 mm Hg) was instituted. The surgery was done in the prone position. Intraoperatively, crystalloid (normal saline/Ringer lactate/kabilyte) was given (at the maintenance rate of 4 mL/kg/hr). Anesthesia was maintained with isoflurane (Minimum Alveolar Concentration [MAC] 0.8–1) in an air–oxygen mixture (60:40). The anesthesia was managed using the Dragger Primus anesthesia workstation. All the patients were operated by two surgeons (M.G. and J.S.G.) by the same unilateral microdiscectomy technique. In this technique, minimal unilateral subperiosteal dissection, laminotomy, and medial third facet drilling were done. Minimal cautery was used in both the groups especially for coagulating any bleeder during subcutaneous and subperiosteal dissection.

Randomization

The participants were randomized into two groups (group M and group D) by a computer-generated random number table. The participant allotment was concealed using sequentially numbered sealed opaque envelopes. A separate investigator opened the sealed envelope intraoperatively prior to soaking the gelfoam with the appropriate test drug. He/she was not involved in any other part of the study. The participant, the surgeon who implanted the gelfoam soaked with the study drug in the epidural space, the intraoperative personnel, the data analyst, and the investigator who monitored the participant in the postoperative period, were blinded to group allocation

At the end of the surgery, gelfoam soaked with the study drug was placed over the exposed duramater, suction was not used locally after this. In group M, 3 mg preservative-free morphine and in group D 8 mg dexamethasone was used to soak the gelfoam placed over the exposed duramater. Intravenous fentanyl infusion was stopped and all patients received intravenous 1 g paracetamol and 4 mg ondansetron during skin closure. The participant was then placed supine and the trachea was extubated after the return of adequate muscle strength and consciousness level. The participants were shifted to the postanesthesia care unit (PACU) for subsequent management.

The participants were assessed in the PACU for level of sedation by using Ramsay sedation score. Each episode of pain was recorded by a blinded observer by using the linear VAS (0–10 cm) scale at 0, 6, 12, 18, and 24 hours in the postoperative period as described during preoperative visit. Participants were asked to make a mark on the line which represents their level of perceived pain intensity. Rescue analgesic (intravenous paracetamol 1 g) was given for VAS ≥ 4 or at participant's request. The time to receive the first rescue analgesic and the total dose of analgesic was recorded. Complete response is defined as minimal to no pain perception by the participant (VAS < 4) in the absence of rescue analgesic medication during the study period (first 24 postoperative hours). Any adverse events such as episodes of desaturation (SpO2 < 92%), hemodynamic disturbances, nausea, and vomiting were noted. The participants were observed for wound healing, infection, hyperglycemia, and opioid-related side effects like itching, urinary retention, constipation, etc. The patients were mobilized from evening of the surgery or the morning of day 1 of surgery. We usually discharge our patients next day of surgery or on day 2.

Statistical Analysis

The sample size of the study was calculated based on the study by Kundra et al. ¹⁴ The mean time to first rescue analgesic was 30.03 ± 6.79 hours in the morphine group in their study. To demonstrate and assumed mean difference of 5 hours, with 5% level of significance, 80% power of study, the sample size came to be 30 in each group. The primary outcome was analyzed using a two-tailed unpaired t -test as well as a Cox proportional hazards model. The categorical outcome of complete response was compared using a chi-squared test. The other numerical outcomes were compared using a linear mixed effects model accounting for repeated measures and the categorical repeated measure outcome (number needing rescue analgesia) was evaluated using a binomial generalized estimating equations model. Post hoc significance tests for pairwise comparisons were done with Bonferroni corrected p -values. The baseline imbalance was assessed using standardized differences. If the 95% confidence interval [CI] of the standardized difference did not include zero then the variable was considered as imbalanced and has been used to provide adjusted estimates (Cox proportional hazards model). A p -value of less than 0.05 has been considered statistically significant. All analyses were performed using R (R Core Team (2023), R Foundation for Statistical Computing, Vienna, Austria).

Results

The CONSORT diagram of participants' enrolment and analysis is shown in Fig. 1 . Seventy participants were assessed for eligibility, 10 participants did not meet the inclusion criteria, thus 60 participants were randomized (30 in group M and 30 participants in group D).

Fig. 1.

CONSORT flow diagram of the study.

The demographic characteristics of the participants (age, gender, American society of Anesthesiologist status, body mass index) were comparable in both the groups. The preoperative hemodynamics (heart rate, systolic and diastolic blood pressure, and oxygen saturation, SpO ₂ ) were also comparable between both the groups. There was no difference between the total intraoperative fentanyl consumption between both the groups. The duration of anesthesia was higher by a mean difference of 46 minutes in group M ( Table 1 ).

Table 1. The baseline demographic characteristics of the patients.

Characteristic		Group D, N  = 30 a	Group M, N  = 30 a	Standardized difference b
Age in years		43.40 (13.65)	46.57 (10.82)	−0.26 [−0.77, 0.25]
Female gender		15.00 (50.00%)	15.00 (50.00%)	0.00 [−0.51, 0.51]
ASA	I	23.00 (76.67%)	21.00 (70.00%)
	II	6.00 (20.00%)	9.00 (30.00%)
	III	1.00 (3.33%)	0.00 (0.00%)
Body mass index, kg/m 2		23.92 (4.46)	24.15 (4.60)	−0.05 [−0.56, 0.45]
Duration of anesthesia, min		181.97 (68.89)	227.90 (96.24)	−0.56 [−1.1, −0.04]
Heart rate, beats/min		84.50 (9.86)	78.67 (10.59)	0.58 [0.06, 1.1]
Systolic blood pressure, mm Hg		128.20 (15.75)	125.90 (16.25)	0.15 [−0.36, 0.65]
Diastolic blood pressure, mm Hg		78.97 (8.03)	76.37 (9.17)	0.31–0.20, 0.82]
SpO2, %		99.10 (1.03)	99.57 (0.73)	−0.53 [−1.0, −0.02]
Total intraoperative fentanyl, μg		183.33 (44.20)	195.83 (52.56)	−0.26 [−0.77, 0.25]

Abbreviations: ASA, American Society of Anesthesiologist; D, dexamethasone; M, morphine; SD, standard deviation; SpO2, saturation of oxygen in peripheral blood.

^a Mean (SD), n (%).

^b95% confidence interval.

There was no statistically significant difference in the mean time to first analgesic in hours in both the groups (M vs. D, 26 [12] vs. 19 [6], mean difference = 7 [95% CI, −8.2, 22, p  = 0.3, unpaired t -test). The hazard ratio of a group M participant receiving an analgesic in the first 24 hours was 0.31 (95% CI, 0.11, 0.86, p  = 0.025) when assessed using a Cox proportional hazards model (adjusting for the duration of anesthesia this was 0.34 [95% CI 0.12–0.99], p  = 0.049).

The pain scores showed a tendency toward low pain scores at 24 hours in the morphine group (group M) ( Table 2 ). The sedation scores and number needing rescue analgesia did not differ between groups except for a tendency for a lower number needing rescue analgesia at 18 hours in group M. The VAS scores and the number needing rescue analgesia decreased significantly (statistically significant time effect, Table 2 ) in both groups over time. There was a difference in terms of complete response between the groups (M vs. D, 83% vs. 57%, difference of 27% [95% CI of 1.1%, 52%], p  = 0.02, chi-square test) ( Fig. 2 , Table 2 ). There was minimal difference in the total grams of paracetamol received as rescue analgesia (M vs. D, 1.00 [0.00] vs. 1.46 [0.78], mean difference = −0.46 [95% CI, −0.93, 0.01, p  = 0.053, unpaired t -test).

Table 2. Comparison of VAS, sedation scores, and number needing rescue analgesia between both the groups at various time points.

Characteristic	Time point	D, N  = 30 (50%) a	M, N  = 30 (50%) a	Difference b (group M – group D)	p -Value c
					Pairwise comparisons d	Group effect	Time effect
Pain scores (VAS)	0 h	2.00 (1.00, 3.00)	2.00 (1.00, 3.00)	0 [−0.94, 0.94]	> 0.9	0.004	< 0.001
	6 h	2.00 (1.00, 2.00)	2.00 (2.00, 2.00)	0 [−0.66, 0.66]	0.6
	12 h	2.00 (2.00, 3.00)	2.00 (2.00, 3.00)	0 [−0.94, 0.94]	> 0.9
	18 h	2.00 (1.00, 5.75)	2.00 (1.00, 3.00)	0 [−2.74, 2.74]	0.3
	24 h	3.00 (2.00, 3.00)	2.00 (1.25, 2.00)	−1 [−1.66, −0.34]	0.11
Sedation score	0 h	2.00 (2.00, 2.00)	2.00 (2.00, 3.00)	0 [−0.66, 0.66]	> 0.9	0.064	0.663
	6 h	2.00 (2.00, 2.00)	2.00 (2.00, 2.00)	0 [0, 0]	> 0.9
	12 h	2.00 (2.00, 2.00)	2.00 (2.00, 2.00)	0 [0, 0]	> 0.9
	18 h	2.00 (2.00, 2.00)	2.00 (2.00, 2.00)	0 [0, 0]	> 0.9
	24 h	2.00 (2.00, 2.00)	2.00 (2.00, 2.00)	0 [0, 0]	> 0.9
Number needing rescue analgesia	0 h	0.00 (0.00%)	0.00 (0.00%)	0% [0%, 0%]	> 0.9	0.912	< 0.001
	6 h	0.00 (0.00%)	0.00 (0.00%)	0% [0%, 0%]	> 0.9
	12 h	2.00 (6.67%)	0.00 (0.00%)	−6.7% [19%, −5.6%]	0.8
	18 h	10.00 (33.33%)	2.00 (6.67%)	−27% [−4.2%, −49%]	0.049
	24 h	4.00 (13.33%)	2.00 (6.67%)	−6.7% [25%, −12%]	0.8

Abbreviations: IQR, interquartile range; VAS, visual analog scale.

^a Median (IQR); n (%).

^bDifference in medians or proportions (group M – group D) [95% confidence interval].

^cLinear mixed effects model/generalized estimating equations (binomial).

^d Bonferroni adjusted p -values.

Fig. 2.

Postoperative outcomes showing ( A ) boxplot of postoperative pain scores as assessed using a visual analog scale (VAS) over time; ( B ) number needing rescue analgesia over postoperative hours; ( C ) boxplot of postoperative sedation scores; and ( D ) proportion with complete response (green) in each group. The dots in ( A ) and ( C ) represents individual data points in each group.

The incidence of side effects was low in both groups. Only one patient in group M had nausea in the PACU. There were no episodes of respiratory depression, pruritus, urinary retention, vomiting bradycardia, or local wound infection in either group for 24 hours.

Discussion

In our study, the mean time to first rescue analgesia following the implantation of gelfoam-soaked epidural morphine was longer (26 hours) as compared with gelfoam-soaked epidural dexamethasone (19 hours). However, this difference was not statistically significant (mean difference = 7 [95% CI, −8.2, 22], p -value = 0.3). The CI of the mean difference is not centered on zero while including practically significant values. The lower end of this CI (−8.2) is quite close to zero relative to the farther upper end of 22 hours indicating the possibility that larger trials could demonstrate a statistically significant difference. Kundra et al ¹⁴ reported that epidural morphine provided a mean time to first rescue analgesia of 30.03 hours in lumbar laminectomy patients. The lower mean time to first rescue analgesia in group M in our study (26 hours) may be due to the lower dose of morphine used (3 vs. 5 mg), which could explain the reasons why the statistical significance was not achieved in the primary outcome in our study.

However, the hazard ratio of the participants in the morphine group requiring rescue analgesia in the first 24 hours was 0.31 (95% CI, 0.11, 0.86, p  = 0.025). This suggests that the patients in the morphine group had 69% less risk of requiring analgesia within first 24 hours compared with the dexamethasone group supporting the superiority of morphine for prolonged postoperative pain relief. A meta-analysis by Wilson-Smith et al ¹³ demonstrated that perioperative epidural steroids significantly decreased postoperative VAS at 24 hours and reduced the number of patients requiring rescue analgesia. While our study confirms the analgesic benefits of dexamethasone, the morphine group showed even greater efficacy in terms of complete response (83% vs. 57%, p  = 0.02).

Gelfoam-soaked epidural morphine had lower VAS score for pain at 24 hours as compared with gelfoam-soaked epidural dexamethasone ( Table 2 ). This statistically significant difference in the pain VAS score between both the groups indicates that morphine provides superior analgesia over an extended period (24 hours) compared with dexamethasone. Dualé et al ¹⁵ compared epidural morphine (2 mg) with intrathecal morphine (0.075 mg) as adjuvant to bupivacaine for postoperative analgesia in cesarean section and concluded that epidural morphine provides longer analgesia than intrathecal morphine. Kumari et al ¹² concluded that epidural gelfoam-soaked levobupivacaine with dexamethasone prolonged postoperative analgesia after lumbar laminectomy, but their study combined dexamethasone with local anesthesia, which confounds neurological assessment.

Gelfoam when implanted in sites without major bleeding liquefies in 7 days and may last up to 4 to 7 weeks. ¹⁶ Adding a drug to this gelfoam should theoretically act as a depot preparation with slow local diffusion of the drug. The morphine group showed a tendency toward lower VAS scores at 24 hours. This is possibly because of the slow release of the drug from the gelfoam, which prolongs the analgesic effect of morphine known to act the spinal cord level. ¹⁷ Epidural dexamethasone may reduce pain by reducing tissue inflammation, by altering transmission in the C fibers or by suppressing prostaglandin formation in the spinal cord. ^{18 19} It is possible that the rate of release from the gelfoam could be different for both morphine and dexamethasone. Consequently, the concentration required to achieve analgesia would be different at different time points between the two drugs and this could explain the lower hazard of requiring rescue analgesia in the morphine group.

Morphine is known to have sedative effects. Our study results showed no significant difference between the morphine group and dexamethasone group in terms of sedation scores at various time points till 24 postoperative hours. The lack of sedation effect could be explained by the lower dose of morphine (3 mg) used in our study.

In our study, the number needing rescue analgesia showed no significant difference between both the groups at various time points (0, 6, 12, 24 hours) except at 18 hours (2 patients in group M vs. 10 patients in group D, p  = 0.049). The results of this study also demonstrated that the number of patients requiring rescue analgesia decreased significantly in both the groups over time, thus, suggesting that with time both the drugs were effective in providing pain relief to the extent that decreased the requirement of recue analgesia. However, the patients in group M had higher rate of complete response (analgesia) as compared with group D (83% vs. 57%, p  = 0.02) highlighting the superiority of morphine in achieving better analgesia than dexamethasone.

The total amount of paracetamol (rescue analgesia) consumption in grams was higher in group D but was not statistically significant. Our study evaluates the use of epidural morphine and dexamethasone in isolation with local anesthetics admixture as local anesthetics produce motor and sensory blockade confounding the immediate postoperative assessment of motor and sensory function of the lower limbs following lumbar spine surgeries. A study by Aljabi et al ²⁰ concluded that absorbable collagen sponge soaked in 80 mg methylprednisolone significantly reduces postoperative pain and length of hospital stay as compared with absorbable collagen sponge soaked in saline in the patients undergoing lumbar microdiscectomy surgeries. We preferred dexamethasone instead of methylprednisolone in our study as it is cheaper and more readily available in our setting.

Morphine induces analgesia by acting on “µ” receptors on the dorsal horn of spinal cord (instead of spinal nerves roots), thus proving analgesia without sympathetic or motor blockade. Its epidural administration has 5 to 10 times more potency than its intravenous administration. A study by Sun et al ²¹ showed that the epidural morphine in a dose of 3 mg provides equal efficacy and lower incidence of side effects when compared with 6 mg dose. Hence, we preferred 3 mg epidural morphine to minimize its side effects in our study. Dexamethasone has been used in a dose of 4 to 8 mg epidurally for providing analgesia. ²² A study by Hefni et al ²³ concluded that the dose of 8 mg provides much superior analgesia as compared with 4 or 6 mg. In our study, we used dexamethasone in the dose of 8 mg.

Sustained and slow release of morphine by the gelfoam might be the cause of no to minimal side effects in the morphine group in our study. Only one patient experienced nausea in our study. This is concurrent with the study by Kundra et al, ¹⁴ which concluded that 5 mg of gelfoam-soaked epidural morphine was associated with minimal side effects (pruritus, nausea, vomiting).

The strength of our study is that it is a prospective randomized clinical trial exploring the efficacy of an emerging technique in using gelfoam as a depot for prolonged analgesic drug delivery. Our study had few limitations such as being a single-center study with a small sample size. Another limitation is that the assessment of pain was done using VAS, which is a subjective score. Moreover, we also did not evaluate dynamic VAS for pain associated with movement. Our findings may be different in other spine surgeries involving vertebral levels. A multicentric study, with large sample, might be required to further validate our results.

Conclusion

Gelfoam-soaked epidural morphine and gelfoam-soaked epidural dexamethasone are safe and appear to have no difference in terms of the time to first rescue analgesic following lumbar discectomy surgeries. However, gelfoam-soaked epidural morphine appears to reduce the hazard of requiring rescue analgesia in the first 24 postoperative hours compared with gelfoam-soaked epidural dexamethasone.