A Randomized, Double-blind, Non-inferiority Trial of Magnesium Sulphate versus Dexamethasone for Prevention of Postoperative Sore Throat after Lumbar Spinal Surgery in the Prone Position.

BACKGROUND: Postoperative sore throat (POST) is a frequent complication of tracheal intubation, particularly after surgery in the prone position. We designed this study to validate the non-inferiority of magnesium sulphate against dexamethasone for prevention of POST after lumbar spinal surgery.
METHODS: One hundred and forty-six patients were randomly allocated to receive either magnesium or dexamethasone. Before anesthetic induction, the magnesium group (n = 73) received magnesium sulphate 30 mg/kg followed by 10 mg/kg/h by continuous infusion until the end of surgery. The dexamethasone group (n = 73) received dexamethasone 8 mg. The primary endpoint was the overall incidence of POST, which was assessed serially over 48 hr postoperatively. The predefined margin of non-inferiority for magnesium against dexamethasone was 15%.
RESULTS: Overall incidences of POST at rest (50.7% versus 49.3% in the magnesium and dexamethasone group, respectively, p = 0.869) and swallowing (65.8% versus 61.6% in the magnesium and dexamethasone group, respectively, p = 0.606) were not different between the groups. The upper limit of the 90% confidence interval, which must be lower than the predefined margin of non-inferiority to prove the non-inferiority of magnesium sulphate against dexamethasone, for at rest and swallowing were 14.97% (p = 0.0496) and 17.19% (p = 0.0854), respectively. The incidences and severities of POST and hoarseness were also not different between the groups throughout the study period.
CONCLUSIONS: Prophylactic magnesium sulphate appears to be non-inferior to dexamethasone for the prevention of POST at rest in patients undergoing lumbar spinal surgery in the prone position.

RCT Entities:   Population Interventions Outcomes

Introduction

Despite being perceived as a minor complication, postoperative sore throat (POST) is common after general anesthesia, occurring in 30% to 65% of patients 1,2. POST contributes to patient dissatisfaction and physical discomfort during the recovery phase 3. POST is attributed to multiple perioperative conditions, including local irritation and inflammation of airway 4, large size of the tracheal tube 5, increased duration of surgery 6, movement of tracheal tube and cuff during position change 1, airway damage after intubation using rigid stylet 1,7, and prone position 8.

In lumbar spinal surgery performed in the prone position, the incidence and severity of POST is expected to be high and of great magnitude as it mandates the use of wire-reinforced tracheal tube having larger outer diameter 9 requiring concomitant use of a rigid stylet, and two times of position change while being intubated. Thus, efforts to reduce the occurrence of POST would be of high priority in this subset of patients while evidence is scarce in that regard.

Numerous pharmacological and non-pharmacological approaches to prevent or minimize the incidence and severity of POST have been tried with conflicting results 4,10,11. Among them, dexamethasone possessing pluripotent action mechanisms has been proposed as a relatively well-proven preventive measure for POST 2,12, although possible inherent complications such as gastric ulcer, infection and hyperglycemia limit its widespread use 13.

Magnesium, an N-methyl-D-aspartate (NMDA) receptor antagonist, has potential advantages to play a beneficial role in reducing POST through multiple mechanisms as well, such as anti-nociceptive and anti-inflammatory effects by inhibiting NMDA receptor-mediated calcium influx 14,15. Indeed, there are evidences to suggest that magnesium can be used for reducing the incidence of POST 16,17. The effect of magnesium on POST, however, has rarely been compared with other agents shown to have efficacies in that regard, especially in patients undergoing lumbar spinal surgery in the prone position who are at increased risk of developing POST.

Therefore, we designed a non-inferior randomized controlled trial to compare the efficacy of magnesium sulphate with dexamethasone for reducing POST in patients undergoing lumbar spinal surgery in the prone position.

Methods

Ethics

This prospective, randomized, double-blind non-inferiority study was conducted at Yonsei University Health System, Seoul, Republic of Korea. Ethical approval for this study (IRB number, 4-2013-0385) was provided by the Institutional Review Board of Yonsei University Health System, Seoul, Republic of Korea. This study was registered at www.ClinicalTrials.gov (identified: NCT01923831). Informed consents were obtained from all patients before enrollment in the study.

Patient selection

After obtaining informed written consent from all patients, 146 patients requiring general anesthesia with tracheal intubation in the prone position for elective lumbar spinal surgery, using standard posterior approach, were enrolled. Inclusion criteria were 20 to 65 year old patients with ASA physical status I-II. Exclusion criteria were pre-existing hoarseness or sore throat, upper respiratory tract infection, use of corticosteroids and calcium channel blockers, allergy to magnesium sulphate, Cormack-Lehane grade 3 or 4, anticipated difficult intubation (a Mallampati score III or IV, underlying facial anomaly or a mouth opening less than 40mm), more than two attempts at intubation, impaired renal function, diabetes mellitus or immunocompromized patients.

Study design

Patients were randomly allocated into either magnesium group (n=73) or dexamethasone group (n=73) using a computerized randomization table the day before operation by an investigator (J.J), not involved with patient care. For each patient, group assignments were kept in sequentially numbered opaque, sealed envelopes that were opened by an investigator (J.H.K) not involved with patient care.

In the magnesium group, magnesium sulphate 30 mg/kg in a total of 100 ml normal saline was given for 10 min before the induction of anesthesia, followed by continuous infusion of magnesium sulphate at 10 mg/kg/h until the end of operation. The chosen dose of magnesium sulphate in this study was determined according to a previous study reporting efficacy of magnesium for pain control with minimal hemodynamic derangement 18. In the dexamethasone group, dexamethasone 8 mg in 100 ml normal saline was given for 10 min before the induction of anesthesia, followed by continuous infusion of equivalent volume of 0.9% saline as in the magnesium group. Infusion of magnesium sulphate or normal saline was discontinued at skin closure. Study drugs were prepared in syringes labelled 'study drug' by an anesthesiologist, not involved in the study.

Perioperative management

Anesthesia was provided by an anesthesiologist (J.K.S), blinded to the patients' group allocation. Upon arrival at the operating room, standard monitoring devices were applied. General anesthesia was induced with propofol 1.5 mg/kg and remifentanil 0.3-0.5 μg/kg/min, and rocuronium 0.6 mg/kg was used for neuromuscular blockade. After loss of all four twitches from train-of-four (TOF) stimulation of the ulnar nerve, direct laryngoscopy was performed using either a Macintosh 3 or 4 laryngoscope blade. Wire-reinforced tracheal tubes (MallinckrodtTM Lo-Contour cuffed reinforced oral/nasal tracheal tube, CovidienTM, Mansfield, OH, USA) of size 8.0 mm and 7.0 mm internal diameter were inserted for male and female patients, respectively, with the aid of a rigid stylet. Application of external laryngeal pressure to aid tracheal intubation, and Cormack and Lehane grade were recorded. All intubation procedures were performed by an anesthesiologist with an at least 2 years of experience, not involved in this study. The intra-cuff pressure was monitored with a noninvasive manometer every 15 min and adjusted between 20-25 cmH2O during operation.

Anesthesia was maintained with sevoflurane 1.5-2.5% in an air-oxygen mixture (fraction of inspired oxygen = 0.5) and 0.1-0.3 μg/kg/min remifentanil infusion. Neuromuscular blockade was monitored using TOF stimulation of ulnar nerve. If hypotension occurred, reduction in mean arterial pressure below 20% of preanesthetic value, 200 ml of either Hartmann's solution or normal saline was given. If hypotension persisted for more than 10 minutes, ephedrine bolus (4mg, repeated) was administered and recorded. If hypotension could not be treated by incremental doses of ephedrine, either phenylephrine or noradrenaline infusion was started. In all patients, prone positioning was done using the Wilson frame with the subject's face resting on a foam cushion (Gentle Touch Headrest Pillow with Intubation Slot, Mizuho OSI, Tokyo, Japan) without rotation of the head.

At the beginning of the skin closure, an i.v. patient-controlled analgesia (PCA) device (Accufuser plus ®, P2015M, Woo Young Medical Co. Ltd, Korea) containing fentanyl 20 μg/kg and ramosetron 0.3 mg in normal saline with a total volume of 100 ml was connected. This was programmed to deliver 2 mL/h as background infusion with a 0.5 ml bolus at 15 min lockout period. At the end of the surgery, oxygen 100% was administered and residual neuromuscular blockade was antagonized with pyridostigmine 0.3 mg/kg and glycopyrrolate 0.004 mg/kg when the four responses were detected in TOF stimulation. Tracheal tube was removed after patients were fully awake. Duration from anesthetic discontinuation to response to verbal command and tracheal extubation were recorded. The patients were allowed to receive tramadol 1mg/kg i.v. if they complained of pain >5 cm on visual analog scale (VAS) score in recovery room and ward. Total number of additional rescue analgesics on recovery room and ward for postoperative analgesia was recorded.

Outcomes

Assessed primary efficacy variables were the overall incidence and severity of POST during the study period. The variables were assessed by an anesthesiologist (J.W.S), blinded to the group allocation, using Harding's direct questionnaire 19. The intensity of sore throat was assessed at rest and on swallowing using a 0-10 cm VAS where a score of 0 meant no pain and a score of 10 represented the worst pain imaginable.

Assessed secondary efficacy variables were intensity of hoarseness using a 4-graded scale [0 = no hoarseness, 1 = mild hoarseness (noticed by the patient only), 2 = severe hoarseness (noticed at the time of the interview by the personnel), 3 = aphonia (inability to speak)], and the following variables: VAS scores of wound pain at rest and upon moving, additional requirement of rescue analgesics, episodes of postoperative nausea and vomiting (PONV). These variables were also assessed by an anesthesiologist (J.W.S).

All above mentioned variables were assessed serially at 4 periods: 0 to 1, 1 to 6, 6 to 24, and 24 to 48 hours after operation. Any complication associated with magnesium sulphate (e.g., diarrhea, hypotension) and dexamethasone (e.g., poor glycemic control) was also assessed.

This study was conducted without any changes of methods or trial outcomes after trial commencement. An anesthesiologist (J.H.P) blinded to group allocation was responsible for collecting data except primary and secondary efficacy variables.

Sample size calculation

The institutional review board highly recommended not designating a control group in this study design, for it would be unethical to neglect the increased risk of POST in the control group. Therefore, this study was performed as a non-inferiority trial, designating dexamethasone, the commonly used drug for preventing POST, as the comparison for the effects of magnesium toward POST. As POST had not been addressed in this studied patients, we used the highest published incidence of POST for sample size calculation. Previous studies have demonstrated that the highest incidence of POST with magnesium and dexamethasone were 23% and 27%, respectively 2,16. The number of patients required for each group was determined on the basis of the non-inferiority hypothesis. For non-inferiority of the magnesium group versus dexamethasone group, a maximum difference of 15 % (margin of non-inferiority) on the incidence of POST was considered as acceptable. Therefore, we calculated that a study with 65 patients per group would have at least 95% power, with one-sided type I error of 0.025. Taking into consideration of the 10 percent dropout rate, we decided to enroll 73 patients in each group.

Statistical analysis

Statistical analyzes were performed using SAS version 9.3 (SAS Institute Inc., Cary NC, USA). Categorical variables were analyzed using chi-square test or Fisher's exact test and defined as number and percentage. Continuous variables were expressed as mean ± SD or median [interquartile range] according to the normality, using Student's t-test or Mann-Whitney U test, as appropriate. P values of <0.05 were considered significant. As secondary analysis, we used Bonferroni correction to compare all pairs of four evaluation time-points, 0 to 1, 1 to 6, 6 to 24, 24 to 48 hours after operation, for multiple comparisons between the groups and adjusted P value of <0.0125 were considered significant.

Results

Baseline demographic data

Among the 180 patients assessed for eligibility, 146 patients who met the inclusion criteria and agreed to participate in this study were enrolled between Aug 2013 and Sep 2014 (Figure 1). Of the 146 patients, study could be completed in all patients and thus, no patient was excluded from statistical analysis.

Figure 1

Flow chart describing recruitment, allocation, follow-up and analysis

The patients' characteristics and operative data were not different between the groups (Table 1). Duration from anesthetic discontinuation to response to verbal command and tracheal extubation, and the amount of remifentanil infused were similar between the groups. None of the patients in the magnesium group demonstrated signs of prolonged neuromuscular blockade.

Table 1

Baseline characteristics of surgical and operative data of patients

Variables	Magnesium(n = 73)	Dexamethasone(n = 73)	p value
Age (yrs)	51 ± 12	51 ± 12	0.852
Female (n)	32 (44%)	31 (43%)	0.999
Height (cm)	165 ± 10	166 ± 10	0.442
Weight (kg)	67 ± 12	67 ± 19	0.965
ASA (I/II)	26/ 47	33/ 40	0.238
Hypertension (n)	23 (32%)	14 (19%)	0.087
Cerebrovascular accident (n)	3 (4%)	4 (6%)	0.999
Smoking (non/past/current)	40/ 14/ 19	52/ 13/ 8	0.115
Diagnosis (spinal stenosis/disc herniation)	37/ 36	38/ 35	0.868
Surgery (posterior fusion/laminectomy)	30/ 43	36/ 37	0.318
Duration of tracheal intubation (min)	140 [100-200]	130 [86-187]	0.371
Duration of surgery (min)	190 [140-243]	170 [138-245]	0.300
Cormack-Lehane grade (1/2)	49/ 24	58/ 15	0.092
External compression (n)	23 (32%)	14 (19%)	0.087
Time to response to verbal command (min)	10 [6-13]	10 [7-12]	0.831
Time to extubation (min)	11 [9-14]	11 [9-13]	0.562
Total amount of rocuronium (mg)	39 [35-44]	40 [34-46]	0.925

Values are number of patients (percentage), mean ± standard deviation or median [interquartile range].

ASA, American Society of Anesthesiologists Physical Status Classification.

Efficacy data

Primary outcome

The incidence and severity of POST are shown in Table 2. Overall incidences of POST at rest (50.7% versus 49.3% in the magnesium group and dexamethasone group, respectively, p = 0.869) and swallowing (65.8% versus 61.6% in the magnesium group and dexamethasone group, respectively, p = 0.606) were not significantly different between the groups. Respective incidences of POST, which were serially assessed at various predefined time points were also all similar between the groups. Overall, the upper limit of the 90% confidence interval for overall incidence, which must be lower than the predefined margin of non-inferiority (15%) to prove the non-inferiority of magnesium sulphate against dexamethasone, were 14.97% (p = 0.0496) and 17.19% (p = 0.0854) at rest and swallowing, respectively. The upper limits of the 90% confidence interval of the serially assessed time points of assessment were all below 15% (all p values <0.05) except at 6-24 h on swallowing, which was 18.86% (p = 0.13) (Figure 2). The VAS scores for the severity of POST were all similar between the groups with median values ranging from 1 to 2 at all time points of assessment. The number of patients who experienced POST VAS score of higher than 4 during the study period was 11 (15.1%) and 12 (16.4%) in the magnesium and dexamethasone group, respectively (p = 0.820). The number of patients who indicated POST VAS score 1 at least once during the study period was 39 (53.4%) and 38 (52.1%) in the magnesium group and dexamethasone group, respectively (p = 0.868).

Table 2

Incidence and severity of postoperative sore throat

Evaluation time points	Magnesium(n = 73)	Dexamethasone(n = 73)	p value
Overall incidence
sore throat at rest	37 (50.7%)	36 (49.3%)	0.869
sore throat on swallowing	48 (68.5%)	45 (61.6%)	0.606
0-1 hour (incidence/severity*)†
sore throat at rest	32 (43.8%) / 1 [1-2]	34 (46.6%) / 2 [1-3]	0.739 / 0.021
sore throat on swallowing	43 (58.9%) / 2 [1-2]	43 (58.9%) / 2 [1-3]	0.999 / 0.103
1-6 hour (incidence/severity*)†
sore throat at rest	26 (35.6%) / 1 [1-2]	25 (34.2%) / 2 [1-2]	0.999 / 0.255
sore throat on swallowing	36 (49.3%) / 2 [1-2]	35 (47.9%) / 2 [1-3]	0.868 / 0.164
6-24 hour (incidence/severity*)†
sore throat at rest	13 (17.8%) / 1 [1-3]	10 (13.7%) / 2 [1-3]	0.496 / 0.499
sore throat on swallowing	22 (30.1%) / 1 [1-3]	17 (23.3%) / 2 [1-2.5]	0.350 / 0.367
24-48 hour (incidence/severity*)†
sore throat at rest	7 (9.6%) / 1 [1-1]	4 (5.5%) / 1 [1-1]	0.347 / 0.450
sore throat on swallowing	12 (16.4%) / 1 [1-2.75]	8 (11.0%) / 1 [1-1]	0.336 / 0.243

Values are number of patients (percentage) or median [interquartile range]. *Severity of postoperative sore throat is expressed as visual analog scale from 0 to 10.

†P value was corrected with Bonferroni method for multiple comparisons and <0.0125 was considered to be statistically significant.

Figure 2

Difference in the incidence of postoperative sore throat. *Against the hypothesis of 15% difference between the magnesium group and dexamethasone group. †Overall incidence of postoperative sore throat. CI, confidence interval.

Secondary outcomes

Postoperative hoarseness occurred in more than half of patient in both group. The incidence and intensity of hoarseness were not statistically significantly different during the study period (Table 3).

Table 3

Incidence and severity of hoarseness and the number of patients who required vasoactive drugs

	Magnesium(n = 73)	Dexamethasone(n = 73)	p value
Patients with hoarseness*
Overall incidence	43 (58.9%)	42 (57.5%)	0.867
0-1 hour	35 (47.9%)	36 (49.3%)	0.868
1-6 hour	36 (49.3%)	30 (41.1%)	0.318
6-24 hour	19 (26.0%)	15 (20.5%)	0.433
24-48 hour	12 (16.4%)	6 (8.2%)	0.131
Hoarseness severity† (0/1/2/3)*
0-1 hour	38/31/4/0	37/35/1/0	0.398
1-6 hour	37/33/3/0	43/26/4/0	0.484
6-24 hour	54/18/0/1	58/8/5/2	0.014
24-48 hour	61/10/0/2	67/5/0/1	0.384
Number of patients who required vasoactive drugs
Ephedrine	48 (65.8%)	38 (52.1%)	0.093
Phenylephrine	4 (5.5%)	5 (6.8%)	0.731
Noradrenaline	11 (15.1%)	8 (11.0%)	0.461

Values are number of patients (percentage) or median [interquartile range].

*The adjusted p value <0.0125 was considered significant after Bonferroni correction. †Hoarseness severity, 0 = no hoarseness, 1 = mild hoarseness (noticed by the patient only), 2 = severe hoarseness (noticed at the time of the interview by the personnel), 3 = aphonia (inability to speak).

The VAS scores for postoperative pain were similar between the groups throughout the study period, whereas more patients in the magnesium group required additional rescue analgesics at time periods of 6 to 24 hours [33 (45.2%) versus 18 (24.7%), p = 0.009], and 24 to 48 hours [30 (41.1%) versus 9 (12.3%), p <0.001)] after operation. Total dose of tramadol for postoperative analgesia was significantly higher in the magnesium group (118 [22-228] mg) than in the dexamethasone group (57 [0-95] mg) (p=0.001). The cumulative volumes of consumed opioid-based PCA were not different between the groups throughout the study period (Table 4).

Table 4

Postoperative analgesic profiles

	Magnesium(n = 73)	Dexamethasone(n = 73)	p value
VAS scores of pain (at rest/ at moving)*
0-1 hour	4 [2-6]/ 5 [3-7]	4 [3-6]/ 5 [3-7]	0.731/ 0.791
1-6 hour	3 [2-5]/ 5 [3-7]	3 [3-5]/ 5 [4-7]	0.534/ 0.530
6-24 hour	3 [2-4]/ 4 [3-6]	3 [2-3]/ 4 [3-5]	0.056/ 0.335
24-48 hour	2 [2-3]/ 3 [3-5]	2 [1-3]/ 3 [2-4]	0.398/ 0.134
Number of patients who required analgesics*
Overall	55 (75.3%)	47 (64.4%)	0.149
0-1 hour	26 (35.6%)	20 (27.4%)	0.285
1-6 hour	30 (41.1%)	27 (37.0%)	0.611
6-24 hour	33 (45.2%)	18 (24.7%)	0.009
24-48 hour	30 (41.1%)	9 (12.3%)	<0.001
Total tramadol dose for pain rescue (mg)	118 [22-228]	57 [0-95]	0.001
Cumulative volume of PCA consumed (mL)*
0-1 hour	1 [0-2]	1 [0-2]	0.752
1-6 hour	15 [13.5-20]	15 [14-20]	0.383
6-24 hour	60 [50-62.5]	58 [50-60]	0.652
24-48 hour	100 [95-100]	100 [100-100]	0.216

Values are number of patients (percentage) or median [interquartile range].

VAS, visual analog scale; PCA, patient-controlled analgesia.

*The adjusted p value <0.0125 was considered significant after Bonferroni correction.

Blood glucose level after operation was significantly higher in dexamethasone group (136.3 ± 32.92 mg/dL versus 116.1 ± 21.17 mg/dL, p <0.001) while preoperative glucose level was similar between the groups. Level of maximal blood magnesium concentration was 1.85 [1.745-2.015] mg/dL in the magnesium group and 1.33 [1.290-1.430] mg/dL in the dexamethasone group (p <0.001). The number of patients who experienced at least one hypotensive event was 53 (72.6%) and 40 (54.8%) in the magnesium group and dexamethasone group, respectively (p = 0.003). In all hypotensive events, blood pressure could be easily restored by predefined measures alone. Overall incidence of PONV was 35.6% in the magnesium group and 31.5% in the dexamethasone group (p = 0.599). Eleven patients were treated with antibiotics for postoperative wound infection; 5 patients in the magnesium group and 6 patients in the dexamethasone group (p = 0.754). Diarrhea was present in one patient in the magnesium group at the second postoperative day, which was spontaneously resolved without any intervention (Table 5).

Table 5

Complications associated with the studied drugs

	Magnesium(n = 73)	Dexamethasone(n = 73)	p value
Preoperative glucose (mg/dL)	98.3 ± 13.89	102.0 ± 14.91	0.122
Postoperative glucose (mg/dL)	116.1 ± 21.17	136.3 ± 32.92	<0.001
Maximal magnesium (mg/dL)	1.85 [1.745-2.015]	1.33 [1.290-1.430]	<0.001
Perioperative hypotension	53 (72.6%)	40 (54.8%)	0.003
Vasopressor use	49 (67.1%)	38 (52.1%)	0.064
Postoperative nausea and vomiting	26 (35.6%)	23 (31.5%)	0.599
Postoperative wound infection	5 (6.8%)	6 (8.2%)	0.754
Diarrhea	1 (1.4%)	0 (0.0%)	>0.999

Values are number of patients (percentage), mean ± standard deviation or median [interquartile range].

Discussion

In this prospective, double-blind, randomized, and non-inferiority trial, we could not observe any statistically significant differences in the incidence and severity of POST between the magnesium and dexamethasone group in patients undergoing lumbar spinal surgery in the prone position. The majority of patients in both groups presented POST VAS scores of less than 3. In addition, the incidence and severity of hoarseness were also not different between the groups. Non-inferiority analysis using the upper limit of the 90% confidence interval revealed that magnesium sulphate appears to be non-inferior to dexamethasone for the prevention of POST at rest in this subset of patients.

Albeit the clinicians have perceived POST as a natural consequence of tracheal intubation rather than complication 20, it may significantly affect the quality of life and satisfaction of patients during the recovery period 21. Consequently, increasing efforts for preventing or minimizing the occurrence of POST during the postoperative recovery period have been attempted 1, while no conclusive measures were elucidated yet, especially in patients undergoing lumbar spinal surgery.

Since POST is mainly mediated by an inflammatory response causing airway mucosal trauma or tissue damage 1, and is a sort of postoperative pain frequently aggravated with PONV 6, pharmacologic agents possessing anti-inflammatory, anti-nociceptive, and anti-emetic effects would be ideal for the prevention and treatment of POST. Various agents have been examined for POST such as dexamethasone 2, inhaled beclomethasone 11, benzydamine hydrochloride spray 22, or transdermal ketoprofen with controversial results 23. Among them, dexamethasone showed promising results 24, which might be attributable to its multidisciplinary mechanisms. Magnesium also has a theoretical potential to reduce POST in terms of anti-inflammatory 25 and anti-nociceptive effect 16, though the effect on PONV was controversial. Recently, magnesium showed a beneficial effect on POST at a clinically relevant dose without complications 16. Therefore, we compared the effects of magnesium and dexamethasone on POST in patients undergoing lumbar spinal surgery in the prone position, who are at increased risk of developing POST, which has not been addressed heretofore. Since the studied patients were at increased risk of developing POST, we designed a non-inferiority trial to avoid establishment of placebo-control group.

As our results indicate, the incidence and severity of POST were all similar between the groups at all time points of assessment. Non-inferiority analysis revealed that magnesium sulphate was non-inferior to dexamethasone for the prevention of POST in terms of the overall incidence at rest, whereas it only showed a trend towards being non-inferior in terms of the overall incidence on swallowing (p = 0.0854). However, detailed analysis revealed that magnesium sulphate was non-inferior to dexamethasone at all serially assessed time points of measurement except at 6-24 h on swallowing, which implicates the non-inferiority of magnesium sulphate to dexamethasone regarding the prevention of POST.

Of note, the overall incidence of POST during study period was more than 60% in both groups despite the use of potentially preventive medications, whereas the majority of patients in this study presented very low VAS scores for POST in comparison with previous studies 2,6,12. The median VAS score of 1 for POST implies that the patient did not perceive POST until they were interviewed. Nonetheless, patients were considered as having POST as long as the VAS score was ≥1 in the current study. Therefore, the high incidence of POST in this study should not discourage the use of both studied drugs as most of the patients had a VAS score of less than 3. Indeed, the overall incidence of patients with POST VAS score of higher than 4 was only 15.1% and 16.4% in the magnesium group and dexamethasone group, respectively (p = 0.820). Thus, these results implicate that both magnesium sulphate and dexamethasone could effectively mitigate the development of clinically significant POST in this high-risk group of patients.

Notably, patients treated with dexamethasone required less rescue analgesics during 6 to 24 hours, and 24 to 48 hours after operation with comparable cumulative infused volume of PCA and VAS scores for pain compared with the patients in the magnesium group. Beneficial effect of dexamethasone to reduce postoperative pain and analgesic requirement in various surgical patients 26 including lumbar spinal surgical patients 27 have been well documented. On the contrary, the effect of magnesium sulphate on postoperative pain is still controversial 28,29, while some studies reported better pain relief with intraoperative magnesium sulphate compared to placebo 15,30. Direct comparison between the magnesium sulphate and dexamethasone against postoperative pain has not been addressed yet, and the result of this study demonstrated superior efficacy of dexamethasone over magnesium sulphate in reducing postoperative pain.

The PONV is known to increase the susceptibility of pain 31,32, and is frequently accompanied with POST 6. Among the available anti-emetics, dexamethasone has been extensively studied and widely used for the reduction of PONV 33. The efficacy of magnesium on antiemetic prophylaxis for PONV is not quite satisfactory heretofore 30. In this study, however, both drugs exerted comparable efficacy on the severity as well as the incidence of PONV throughout the study period.

One of the concerns of using dexamethasone is the potential for poor glycemic control. In this study, postoperative glucose level measured at immediate postoperative period was higher in the dexamethasone group. The clinical significance of this minor increase in blood glucose level in the dexamethasone group is beyond the scope of this study to discuss and remains to be proven. Considering the results of the current trial, magnesium sulphate could be a valuable option for the prevention of POST when the use of dexamethasone may not be appropriate for its potential side-effects.

This study is subject to following limitations. Since we excluded patients with expected difficult intubation to control confounding variables, generalization of this result should be exercised with caution. While the dose of magnesium sulphate was determined based on the estimated safe dose without hemodynamic consequences, the dose of magnesium used in this study might not be an optimal dose for POST, which merits further studies in that regard. Likewise, same limitation applies to the chosen dose of dexamethasone in this study.

In conclusion, prophylactic magnesium sulphate appears to be non-inferior to dexamethasone for prevention of POST at rest in patients undergoing lumbar spinal surgery in the prone position.