Secondary Logo

Journal Logo

Clinical Investigations

Mannitol Improves Intraoperative Brain Relaxation in Patients With a Midline Shift Undergoing Supratentorial Tumor Surgery: A Randomized Controlled Trial

Li, Shu MD*; Sun, Huihui MD; Liu, Xiaoyuan MD*; Ren, Xiaohui MD; Hao, Shuyu MD; Zeng, Min MD*; Wang, Dexiang MD*; Dong, Jia MD*; Kan, Qing MD*; Peng, Yuming MD*; Han, Ruquan MD, PhD*

Author Information
Journal of Neurosurgical Anesthesiology: October 2020 - Volume 32 - Issue 4 - p 307-314
doi: 10.1097/ANA.0000000000000585


Satisfactory brain relaxation is closely associated with clinical outcomes after neurosurgery.1,2 Unsatisfactory brain relaxation and increased intracranial pressure (ICP) are common situations during neurosurgical procedures that can lead to difficulty in surgical exposure and operative procedures, especially in patients with preoperative midline shift.3–6 Therefore, measures to achieve optimal intraoperative brain relaxation in patients with midline shift have received increasing attention in clinical studies.

Several factors are associated with brain relaxation, including characteristics of the intracranial lesion, head and neck position, cerebrospinal fluid drainage, electrolyte disturbance, anesthetic agents, arterial carbon dioxide partial pressure (PaCO2) and osmotic therapy.7–11 Preoperative midline shift and perioperative risk factors potentially confound these effects, but these have not been well controlled in previous related studies thus underpowering the conclusions.4–6 Limited conclusions can be drawn with regard to the effects of mannitol on brain relaxation in patients undergoing supratentorial brain tumor resection.

On the basis of the findings of previous studies, we hypothesized that mannitol would improve brain relaxation in patients with midline shift undergoing elective supratentorial brain tumor surgery. The primary endpoint of this study was the proportion of satisfactory brain relaxation. Secondary endpoints were incidence of adequate dural tension, surgical exposure, requirements for rescue therapies for brain swelling, and postoperative complications.



This was a single-center, randomized, parallel, placebo-controlled and double-blinded trial conducted at Beijing Tian Tan Hospital, Capital Medical University. Ethical approval was obtained from the China Ethics Committee of Registering Clinical Trials (reference number: ChiECRCT2013033). The study was registered at (registry number: ChiCTRTRC13003984). Written informed consent was obtained from all participants or their next of kin one day before surgery, before randomization. The trial protocol was published previously.7


Patients scheduled to undergo elective supratentorial tumor resection were screened for eligibility. The inclusion criterion were age between 18 and 60 years, and preoperative brain imaging (computed tomography or magnetic resonance imaging [MRI]) with midline shift confirmed and calculated by a trained and qualified radiologist based on a previously described method and shown in Figure 1.8

Measurement of midline shift.
. BPD is the biparietal diameter; SP the distance from the inner table to the septum pellucidum on the side of the shift. BPD and SP were calculated at the level of the foramen of Monro based on magnetic resonance imaging performed with a T1 sequence or Computed tomography.

Patients with America Society of Anesthesiologists physical status IV to V, Glasgow Coma Scale (GCS) score <13, plasma sodium level <130 mmol/L or >150 mmol/L, renal insufficiency/creatinine clearance <30 mL/kg, heart disease/cardiac ejection fraction <20%, or preoperative cerebrospinal fluid drainage by lumber puncture or ventricle drainage were excluded from the study.

Randomization and Blinding

A computer-generated randomization table prepared by an investigator not involved in the trial to guarantee allocation concealment was used. Patients were randomly assigned following simple randomization to 1 of 4 parallel groups 24 hours before surgery, with an allocation ratio of 1:1:1:1. The randomization codes were sent to the research nurse who was responsible for preparation of the infusion agents in sterile, plastic and opaque containers (labelled “research solution”) as well as the administration of the investigated agents. Anesthesiologists, neurosurgeons, patients, and the investigators who collected and analyzed the data were all blinded to the investigated drug grouping.


Patients were randomly allocated to receive placebo (200 mL of 0.9% saline) or 20% mannitol at doses of 0.7, 1.0, or 1.4 g/kg at the onset of surgical incision, similar to a previous study.5 The investigational drugs were administered intravenously at 600 mL/h via a reliable intravenous catheter (Supplementary Table 1, Supplemental Digital Content 1,

Concomitant Treatment

Concomitant patient care, including routine monitoring, premedication, arterial line placement, and anesthesia induction and maintenance, are described in the previously published protocol.7 In summary, all patients were premedicated with IV midazolam 0.05 mg/kg 15 minutes before induction of anesthesia. After preoxygenation, anesthesia was induced with IV sufentanil 0.2 to 0.3 μg/kg, propofol (target-controlled infusion with an initial plasma concentration of 4 μg/mL), and rocuronium 0.6 mg/kg. Mechanical ventilation with a 60% fraction of inspired oxygen was performed with the following parameters to maintain normocapnia (PaCO2 35 to 40 mm Hg): tidal volume of 6 to 8 mL/kg, respiratory rate of 12 to 15 per minute, I: E ratio of 1:2, and 5 cm H2O of positive end-expiratory pressure. Anesthesia was maintained with propofol (target-controlled infusion with a plasma concentration of 3 to 5 μg/mL) and remifentanil 0.1 to 0.2 μg/kg/min, with supplemental boluses of IV rocuronium for muscle relaxation as required. Anesthesia was adjusted to maintain mean blood pressure within 20% of baseline. After the initial evaluation of brain relaxation, the neurosurgeon, who was blinded to the study group, could request rescue mannitol, hyperventilation, or both to improve brain relaxation based on their assessment.


Preoperative peritumoral edema was evaluated by measuring the shortest straight line between the tumor edema margins on the T2 sequence of MRI according to the Steinhoff classification9 as follows: 0, no signs of edema; I, mild edema, limited to 2 cm; II, moderate edema, >2 cm but limited to the ipsilateral hemisphere; and III, severe edema, extending to the contralateral hemisphere (Supplementary Table 2, Supplemental Digital Content 2, Blood was sampled from the arterial line before and after the initial mannitol administration for measurement of electrolyte and glucose concentrations.

The primary endpoint was the proportion of satisfactory brain relaxation assessed by the neurosurgeon using a 4-point scale upon opening the dura.5,10 Brain relaxation was dichotomized to satisfactory (scores 1 and 2, representing complete and adequate relaxation) or unsatisfactory (scores 3 and 4, representing a firm and bulging brain) (Supplementary Table 3, Supplemental Digital Content 3, Secondary endpoints included the incidence of adequate dural tension, satisfactory surgical exposure classification, request for rescue treatments, and postoperative complications. Dural tension was evaluated using a 3-point scale after bone flap removal, in which 1 and 2 represented very tight and tight, indicating unrelaxed dural tension, and 3 for relaxed dural tension (Supplementary Table 4, Supplemental Digital Content 4, Surgical exposure classification was graded as satisfactory or unsatisfactory. Brain relaxation, dural tension, and surgical exposure were all assessed or recorded by 2 trained and qualified neurosurgeons blinded to the study group. If these 2 neurosurgeons assessed different scores, a third attending neurosurgeons made the final judgment.

Postoperative complications within 7 days, including intracranial hemorrhage, infection, status epilepticus, severe electrolyte disturbance and death, were recorded. Postoperative cerebral edema (also assessed by the Steinhoff classification on T2 sequence of MRI) was assessed by 2 trained researchers blinded to the study group within 7 days of surgery. The kappa coefficient was calculated between their assessments. Patients with severe symptoms including refractory headache, epileptic seizures, hemiplegia, and deteriorating status of consciousness may have received >1 radiologic examination during the first 7 postoperative days, but only the first MRI scan was evaluated for the purposes of this study.

Statistical Analysis

According to the findings of a previous study,5 the proportion of satisfactory brain relaxation (brain relaxation scores 1 and 2) was assessed as 45% in patients with midline shift administered or not administered low-dose mannitol, and the effect size was estimated as 27%. A sample size of 200 was calculated to achieve 85% power to detect a 2-sided significance level (α) of 0.017, with the addition of 20% to account for dropouts.11 Normality was determined with the Shapiro-Wilk test. Continuous and skewed data were reported as medians with interquartile ranges. Categorical data were described as counts (percentages). Normally distributed and continuous variables were compared with analysis of variance, while skewed variables were compared using the Mann-Whitney U test. Categorical data were analyzed with χ2 test or Fisher’s exact test. For multiple comparisons between-groups, we specified a priori that pairwise comparisons would be performed with P<0.0083, with Bonferroni corrections used to denote statistical significance. The primary outcome was compared among 4 groups by the Cochran-Mantel-Haenszel test. The Cochran-Armitage test for post hoc analysis was used for trend analysis related to dose escalations. Satisfactory brain relaxation was analyzed using multivariate logistics regression model taking into account the risk factors including tumor size, location and type, peritumoral edema, and midline shift. The Hosmer and Lemeshow goodness of fit test and the receiver operating characteristic curve analysis were used to determine goodness of fit as well as model discrimination. A significance level of P<0.05 was used to indicate statistical significance. SAS 9.3 (SAS Institute, Cary, NC) for Windows was used for all statistical analyses.


In total, 469 consecutive patients with supratentorial brain tumors scheduled for elective resection at Beijing Tian Tan Hospital, Capital Medical University, were screened for eligibility from January 2017 to June 2017. In total, 265 patients were excluded, and 204 randomized into placebo and 0.7, 1.0, or 1.4 g/kg mannitol groups. Intraoperative brain relaxation could not be assessed in 4 patients due to dural laceration during drilling, and 21 patients could not be evaluated for postoperative cerebral edema evaluation within 7 days due to unstable status. Therefore, 179 were included in the final analysis (Fig. 2).12

Consort flow diagram of patient recruitment.

Patient demographics, past medical history, tumor characteristics, and preoperative laboratory investigations did not differ between the groups (Table 1). After infusion of mannitol, blood glucose, plasma sodium, plasma potassium, urine output, and total output differed among the 4 groups (P<0.05); however, blood glucose and sodium remained within the normal ranges (Table 2). Twenty-eight patients (14%) developed hypokalemia (plasma K+<3.5 mmol/L), but the number of patients with hypokalemia did not differ among groups.

TABLE 1 - Demographics and Baseline Characteristics
Placebo (N=50) Mannitol: 0.7 (N=50) Mannitol: 1.0 (N=50) Mannitol: 1.4 (N=50)
Age (y) 50 (37-57) 50 (36-60) 47.5 (37-57) 53 (42-57)
Female sex (n [%]) 23 (46) 26 (52) 18 (36) 27 (54)
Height (cm) 165 (160-173) 168 (160-173) 165 (159-172) 166 (159-173)
Weight (kg) 66 (60-75) 68 (63-74) 62 (56-74) 64 (55-75)
BMI 24 (23-26) 24 (22-27) 23 (21-25) 23 (21-27)
Medical history (n [%])
 Recurrence 6 (12) 4 (8) 6 (12) 2 (4)
 Diabetes 3 (6) 3 (6) 3 (6) 3 (6)
 Hypertension 7 (14) 7 (14) 5 (10) 10 (20)
Midline shift (mm) 3.0 (2.0-7.3) 3.0 (3.0-7.0) 5.0 (3.0-9.0) 6.0 (3.0-9.3)
Tumor size (mm3) 70 (33-121) 54 (33-85) 60 (30-134) 69 (29-100)
Tumor classification (n [%])
 Glioma 18 (36) 28 (56) 27 (54) 24 (48)
 Meningioma 27 (54) 20 (40) 22 (44) 24 (48)
 Metastatic 1 (2) 0 (0) 0 (0) 2 (4)
 Other 4 (8) 2 (4) 1 (2) 0 (0)
Tumor site (n [%])
 Frontal 18 (36) 23 (46) 22 (44) 21 (42)
 Parietal 12 (24) 7 (14) 10 (20) 5 (10)
 Temporal 9 (18) 13 (26) 13 (26) 17 (34)
 Occipital 4 (8) 3 (6) 4 (8) 6 (12)
 Other 7 (14) 4 (8) 1 (2) 1 (2)
Preoperative edema classification (n [%])
 0: No signs 7 (14) 2 (4) 3 (6) 1 (2)
 I: Mild 26 (52) 25 (50) 22 (44) 19 (38)
 II: Moderate 13 (26) 19 (38) 19 (38) 26 (52)
 III: Severe 4 (8) 4 (8) 6 (12) 4 (8)
Preoperative steroid (n [%]) 18 (36) 28 (56) 27 (54) 24 (48)
Preoperative mannitol (n [%]) 2 (4) 3 (6) 1 (2) 0 (0)
Preoperative laboratory test
 Hb (g/dL) 140 (133-154) 141 (131-159) 141 (130-151) 144 (134-158)
 WBC (X109) 6.5 (5.2-7.1) 6.4 (5.1-8.4) 6.1 (5.1-7.4) 6.9 (5.7-7.8)
 PLT (X109) 230 (196-281) 236 (191-292) 236 (204-256) 234 (201-267)
 BUN (mmol/L) 5.6 (4.4-6.5) 4.9 (4.3-6.1) 5.1 (4.4-6.1) 5.5 (4.5-6.7)
 AST (U/L) 18 (16-21) 20 (16-24) 20 (17-25) 18 (15-21)
 ALT (U/L) 18 (13-26) 21 (13-30) 20 (16-34) 18 (14-27)
 GLU (mmol/L) 5.2 (4.8-5.7) 5.3 (5.7-5.5) 5.0 (4.8-5.5) 5.0 (4.8-5.5)
 Na (mmol/L) 142 (140-143) 141 (140-143) 141 (140-142) 142 (140-143)
 K (mmol/L) 4.0 (3.9-4.3) 4.3 (4.0-4.5) 4.2 (3.9-4.4) 4.1 (3.9-4.4)
Preoperative mannitol was the number of patients receiving preoperative mannitol infusion to reduce intracranial pressure or alleviate symptom. Data are expressed as the medians (IQR) and n (%), unless otherwise stated. The Shapiro-Wilk test and parametric and nonparametric tests were performed as appropriate. P<0.05 was used to denote statistical significance.
ALT indicates alanine transaminase; AST, aspartate transaminase; BMI, body mass index; BUN, blood urea nitrogen; Cl, chlorine; Glu, glucose; Hb, hemoglobin; K, potassium; Na, sodium; PLT, platelets; WBC, white blood cells.

TABLE 2 - Intraoperative Parameters
Placebo (N=50) Mannitol: 0.7 (N=50) Mannitol: 1.0 (N=50) Mannitol: 1.4 (N=50) P
Non-neutral head position 23 (46) 20 (40) 28 (56) 29 (58) 0.229
Before mannitol infusion
 MAP (mm Hg) 85 (76-93) 89 (80-99) 88 (80-98) 87 (80-98) 0.166
 HR (bpm) 59 (52-66) 61 (54-66) 60 (54-68) 60 (53-68) 0.799
 GLU (mmol/L) 5.2 (4.9-5.7) 5.3 (5.0-5.7) 5.0 (4.7-5.4) 5.2 (4.6-5.5) 0.098
 Na (mmol/L) 143 (141-143) 142 (140-143) 142 (141-144) 142 (141-143) 0.584
 K (mmol/L) 3.7 (3.5-4.0) 3.8 (3.6-4.0) 3.7 (3.4-3.9) 3.7 (3.4-3.9) 0.154
 PaCO2 (mm Hg) 34 (32-36) 36 (33-37) 34 (32-37) 35 (33-38) 0.585
After mannitol infusion
 MAP (mm Hg) 78 (73-91) 84 (74-91) 83 (75-89) 82 (76-89) 0.824
 HR (bpm) 57 (51-62) 57 (52-64) 59 (52-66) 62 (53-68) 0.321
 GLU (mmol/L) 5.4 (4.9-5.8) 5.2 (4.7-5.5) 4.8 (4.5-5.1) 4.8 (4.4.-5.3) <0.0001
 Na (mmol/L) 143 (142-144) 139 (137-140) 138 (135-139) 135 (134-136) <0.0001
 K (mmol/L) 3.7 (3.5-4.0) 4.0 (3.7-4.2) 3.9 (3.6-4.1) 4.0 (3.7-4.3) 0.004
 PaCO2 (mm Hg) 34 (32-35) 34 (32-37) 34 (32-36) 34 (33-36) 0.441
Total input (mL) 2600 (2100-3225) 2600 (2087-3400) 2750 (2138-3363) 3050 (2175-3487) 0.621
Total output (mL) 1500 (1195-2135) 2000 (1343-2803) 2275 (1575-2775) 2675 (1900-3408) <0.0001
 Urine volume 1150 (800-1725) 1500 (1100-2225) 1700 (1238-2325) 2100 (1500-2800) <0.0001
 Blood loss 335 (200-600) 300 (200-525) 300 (200-600) 310 (200-623) 0.867
Surgery duration (min) 242 (198-304) 235 (182-291) 230 (179-286) 244 (194-295) 0.772
Anesthesia duration (min) 281 (235-333) 277.5 (218-336) 270 (215-340) 285 (234-340) 0.707
Data were expressed as the medians (IQR) and n (%) unless otherwise stated. Shapiro-Wilk test and parametric and nonparametric tests were used accordingly. P<0.05 was used to denote statistical significance.
Glu indicates glucose; HR, heart rate; K, potassium; MAP, mean arterial pressure; Na, sodium; PaCO2, arterial partial pressure of carbon dioxide; PaO2, arterial partial pressure of oxygen.

The proportion of satisfactory brain relaxation was significantly higher in the 1.0 g/kg (86%) and 1.4 g/kg mannitol (88%) groups compared with the 0.7 g/kg mannitol (44%) and placebo (44%) groups (P<0.0001). Mannitol produced satisfactory brain relaxation in a dose-dependent manner as determined by the Cochran-Armitage trend test (P<0.0001) (Table 3). The multivariate logistics regression model showed that mannitol dose was associated with a 2.81 (95% confidence interval [CI]: 1.97-4.02, P<0.0001) odds of satisfactory brain relaxation adjusted by tumor size and peritumoral edema (Table 4). In addition, tumor size (odds ratio [OR]: 0.994 per 1 mm3, 95% CI: 0.989-0.998, P=0.004) and peritumoral edema (OR: 0.60 per increased edema grade, 95% CI: 0.37-0.97, P=0.038) were also independent risk factors for satisfactory brain relaxation. Preoperative midline shift on the other hand was not associated with satisfactory brain relaxation (OR: 0.94, 95% CI: 0.87-1.04; P=0.112). In addition, patients with peritumoral edema >2 cm but limited to the ipsilateral hemisphere had a 0.16 (95% CI: 0.03-0.78) odds of satisfactory brain relaxation compared with those with no peritumoral edema (P=0.024). The Hosmer-Lemeshow test showed that the multivariant logistic regression model was a good fit (P>0.05), and the area under the receiver operating characteristic curve was 0.799.

TABLE 3 - The Primary Outcome and Secondary Outcomes
Placebo (N=50) Mannitol: 0.7 (N=50) Mannitol: 1.0 (N=50) Mannitol: 1.4 (N=50) Overall P Cochran-Armitage Test P
Primary outcome
 Brain relaxation 22 (44) 22 (44) 43 (86) 44 (88) <0.0001 <0.0001
Secondary outcome
 Relaxed dural tension 16 (32) 20 (40) 34 (68) 32 (64) <0.0001 <0.0001
 Surgical exposure 22 (44) 26 (52) 44 (88) 44 (88) <0.0001 <0.0001
 Additional treatment 41 (82) 36 (72) 32 (64) 25 (50) 0.007 0.0005
 Complications* 24 (48) 14 (28) 16 (32) 28 (56) 0.013 0.368
 Moderate to severe cerebral edema (n/N)† 24/45 (53) 22/43 (51) 24/45 (53) 36/46 (78) 0.025 0.018
Incidences of the primary outcome and secondary outcomes are presented as the n (%) or n/N (%). Results of the between-groups analyses were presented as the risk ratios (RRs) and 99% confidence intervals (CIs). Additional treatment included hyperventilation and extramannitol rescue.
*Postoperative complications included hematoma, status epilepticus, infection and death. No patients died within 1 week postoperatively.
†Twenty-one patients did not get magnetic resonance scan due to unstable status for postoperative cerebral edema evaluation. P<0.05 was used to denote statistical significance.

TABLE 4 - The Estimated Odds Ratios in the Multivariate Logistics Regression Model for the Analysis of Satisfactory Brain Relaxation
95% Confidence Interval
Parameters Odds Ratio P Lower Limit Upper Limit
Mannitol dose (g/kg)* 2.81 <0.0001 1.97 4.02
 Placebo Ref Ref Ref Ref
 0.7 1.14 0.777 0.47 2.73
 1.0 15.8 <0.0001 4.91 51.3
 1.4 18.3 <0.0001 5.66 59.5
Tumor size (mm3)* 0.994 0.004 0.98 0.99
Peritumor edema* 0.60 0.038 0.37 0.97
 0 Ref Ref Ref Ref
 I 0.48 0.345 0.11 2.18
 II 0.16 0.024 0.03 0.78
 III 0.52 0.512 0.07 3.69
P=0.063>0.05 in the Hosmer-Lemeshow test. Area under receiver operating characteristic (ROC) curve was 0.799.
*Mannitol dose, tumor size, and peritumor edema were included in the multivariate logistics model for predicting satisfactory brain relaxation. Ref indicates reference parameter in the subgroup analysis.

The incidence of relaxed dural tension, satisfactory surgical exposure, requirement for rescue treatments and postoperative moderate to severe cerebral edema all differed significantly among groups and in a dose-dependent manner, similar to the primary endpoint (Table 3). The kappa coefficient for the 2 assessments of postoperative edema evaluation was 0.854 (P<0.0001). There was a significant difference in postoperative moderate to severe (Steinhoff classification grade II and III) cerebral edema among the 4 groups (P=0.025), and in a dose-dependent manner (P=0.018). The 1.4 g/kg dose of mannitol significantly increased the risk of moderate to severe cerebral edema compared with 0.7 g/kg (78% vs. 51%, P=0.007). However, 1.0 g/kg mannitol did not increase the risk (P=0.839). The risk of postoperative complications, including hematoma, status epilepticus, and infection, did not increase with mannitol dose (P=0.368). No patient died within 1 week of surgery. Between-group analysis of the primary and secondary outcomes is presented in Supplementary Table 5 (Supplemental Digital Content 5,


This randomized controlled, double-blinded study, showed that mannitol improved brain relaxation, dural tension and surgical exposure, and decreased the requirement of rescue treatments during supratentorial brain tumor surgery in a dose-dependent manner. In addition, tumor size, peritumoral edema, as well as mannitol dose, influenced satisfactory brain relaxation. However, the highest dose of mannitol (1.4 g/kg) increased the risk of postoperative cerebral edema, which should be considered when choosing mannitol dose.

Many factors contribute to brain relaxation. Intravenous anesthetic agents and decreased PaCO2 are associated with reduction of cerebral metabolic rate and cerebral vasoconstriction, respectively, resulting in brain relaxations.2,13–15 A non-neutral position of the neck impedes venous return, increases ICP, and leads to brain swelling, and ultimately to difficult surgical exposure.16,17 Aside from these factors, mass effect attributable to the size, pathology, and location of the tumor, peritumoral edema and the presence of midline shift, are particularly important factors that can affect the adequacy of brain relaxation.18 Quentin et al5 compared the effects of 0.7 g/kg and 1.4 g/kg mannitol in patients undergoing supratentorial tumor surgery and found that the higher dose of mannitol provided better relaxation in patients with midline shift. However, this study did not control the number of patients with midline shift or provide information about the impact of tumor characteristics on the observed effects. Akcil et al6 demonstrated that 0.5 g/kg mannitol could provide adequate brain relaxation, however, they compared with the combinations of furosemide with different doses of mannitol. Seo and colleagues also compared the effect of 4 doses of mannitol (0.25, 0.5, 1.0, 1.5 g/kg) on brain relaxation in patients undergoing supratentorial tumor resections, and found that 1.0 g/kg was the optimal dose for satisfactory brain relaxation with fewest side effects.11 However, unlike our study, that of Seo and colleagues did not quantitatively measure peritumoral edema. We found that 1.0 g/kg and 1.4 g/kg doses of mannitol provided better brain relaxation than 0.7 g/kg mannitol or placebo, which is in accordance with the findings of Seo et al.4 However, we also took the effect of tumor characteristics into consideration and found that tumor size and peritumoral edema, rather than midline shift, was associated with satisfactory brain relaxation. These findings are in contrast to the results of the study by Quentin and colleagues.12 Midline shift is defined by dividing bipatietal diameter by 2 and subtracting the distance from the inner table to the septum pellucidum on the side of the shift. It is determined by tumor size, tumor location and peritumoral edema. Small tumors with no peritumoral edema that are located close to the midline may therefore, exhibit midline shift on imaging, but might not increase ICP or reduce intracerebral compliance. In contrast, tumors with peritumoral edema >2 cm but limited to the ipsilateral hemisphere may reduce intracerebral compliance. The current study found that tumor size but not midline shift, together with the dose of mannitol, is closely associated with satisfactory brain relaxation. Although 0.7 g/kg did not improve brain relaxation compared with placebo, the proportion of relaxed dural tension and surgical exposure were better, and the need for rescue treatments and complications were less.

In the current study, blood glucose and plasma sodium and potassium levels differed among the 4 groups after infusion of mannitol, although most remained within their normal ranges. We assessed postoperative cerebral edema which was not reported in previous studies. Although 21 patients were unable to undergo a postoperative MRI scan, we still found that moderate to severe postoperative brain edema within 7 days after surgery was significantly different among mannitol doses, being highest in 1.4 g/kg mannitol group. Brain tissue resists sustained volumetric reduction and reacts to parenchymal shrinkage by cellular/tissue expansion.19,20 At the moment when the effect of mannitol wears off, rebound edema, mediated by the reuptake of free water and various electrolytes, begins. In the current study, while 1.4 g/kg mannitol provided the highest proportion of satisfactory brain relaxation it was also associated with the greatest postoperative cerebral edema. We assessed cerebral edema within 7 days of surgery, which is within the time window for the appearance of rebound edema.21 The balance between over-shrunken and satisfactorily relaxed brain should be take into account clinical outcomes.2 It would also be incorrect to recommend an optimal mannitol dose based only on the incidence of postoperative cerebral edema without clear clinical correlation between such edema and clinical outcomes. However, in the absence of outcomes data, a mannitol dose of 1.4 g/kg for brain relaxation should be recommended with caution because of the higher risk of rebound edema. Moreover, the risk of rebound edema should also be considered if higher doses of mannitol are repeated.

There are several limitations to our study. First, due to the lack of lactate and chloride results we were unable to calculate osmolality after mannitol infusion, although these data were reported in the study by Seo et al.4 Second, there was no objective measure of the effect of mannitol on brain relaxation, only a subjective evaluation by the neurosurgeons. However, several clinical studies have also used this same evaluation to assess brain relaxation; its reliability and validity is accepted and it is widely applied in clinical practice.4–6,22,23 Third, our study only included patients with supratentorial tumors, so the results cannot be extrapolated to those with infratentorial lesions or severe intracranial hypertension. Fourth, postoperative complications were assessed only within 1 week after surgery, although the time window for assessment was in accordance with the peak incidence of complications and guideline recommendation.24 However, future studies should include a longer follow-up period. Fifth, potential bias in the assessment of the incidence and degree of postoperative edema may have occurred because of exclusion of patients who did not undergo postoperative MRI scan because of their unstable clinical status. Moreover, the impact of postoperative edema on clinical outcomes was not assessed. There was also no record of the postoperative administration of mannitol for the treatment of edema. Finally, this was a single-center study and its results might not be generalizable. The time from skin incision to opening of the dura is likely diverse among individual neurosurgeons and medical centers. In addition, the patients with midline shift are not routinely administered mannitol before surgery in our center, unless they had the symptoms of increased ICP such as headache and vomiting, and this practice may differ in other centers.

In summary, this randomized controlled double-blinded trial found that intraoperative infusion of mannitol improves brain relaxation in a dose-dependent manner during tumor resection in patients with supratentorial brain tumor causing midline shift. The effects of mannitol on brain relaxation are affected by the tumor size and peritumoral edema classification, rather than by midline shift. High-dose mannitol (1.4 g/kg) should be used with caution for intraoperative brain relaxation because of the potential for postoperative cerebral edema. Further multicentre studies are needed to assess the safety and long-term clinical outcomes of mannitol infusion.


The authors thank Dr. Hongqiu Gu (Tian Tan Clinical Trial and Research Center for Stroke, Department of Neurology, Beijing Tian Tan Hospital, Capital Medical University, China, and Dr. Yuanjun Li (Department of Radiology, Beijing Tian Tan Hospital, Capital Medical University, China, for their help with the statistical plan and radiology support.


1. Joseph B, Friese RS, Sadoun M, et al. The BIG (brain injury guidelines) project: defining the management of traumatic brain injury by acute care surgeons. J Trauma Acute Care Surg. 2014;76:965–969.
2. Li J, Gelb AW, Flexman AM, et al. Definition, evaluation, and management of brain relaxation during craniotomy. Br J Anaesth. 2016;116:759–769.
3. Grandhi R, Bonfield CM, Newman WC, et al. Surgical management of traumatic brain injury: a review of guidelines, pathophysiology, neurophysiology, outcomes, and controversies. J Neurosurg Sci. 2014;58:249–259.
4. Seo H, Kim E, Jung H, et al. A prospective randomized trial of the optimal dose of mannitol for intraoperative brain relaxation in patients undergoing craniotomy for supratentorial brain tumor resection. J Neurosurg. 2017;126:1839–1846.
5. Quentin C, Charbonneau S, Moumdjian R, et al. A comparison of two doses of mannitol on brain relaxation during supratentorial brain tumor craniotomy: a randomized trial. Anesth Analg. 2013;116:862–868.
6. Akcil EF, Dilmen OK, Karabulut ES, et al. Effective and safe mannitol administration in patients undergoing supratentorial tumor surgery: a prospective, randomized and double blind study. Clin Neurol Neurosurg. 2017;159:55–61.
7. Peng Y, Liu X, Wang A, et al. The effect of mannitol on intraoperative brain relaxation in patients undergoing supratentorial tumor surgery: study protocol for a randomized controlled trial. Trials. 2014;15:165.
8. Greenberg MS. Handbook of Neurosurgery. New York: Thieme; 2016.
9. Kazner E, Lanksch W, Steinhoff H, et al. Computerized axial tomography of the skull—diagnostic possibilities and clinical results (author’s transl). Fortschr Neurol Psychiatr Grenzgeb. 1975;43:487–574.
10. Malik ZA, Mir SA, Naqash IA, et al. A prospective, randomized, double blind study to compare the effects of equiosmolar solutions of 3% hypertonic saline and 20% mannitol on reduction of brain-bulk during elective craniotomy for supratentorial brain tumor resection. Anesth Essays Res. 2014;8:388–392.
11. Hulley SB, Cummings SR, Browner WS, et al. Designing Clinical Research: an Epidemiologic Approach. Philadelphia, PA: Lippincott Williams & Wilkins; 2013.
12. Schulz KF, Altman DG, Moher D. CONSORT 2010 statement: updated guidelines for reporting parallel group randomised trials. BMJ. 2010;340:c332.
13. Dinsmore J. Anaesthesia for elective neurosurgery. Br J Anaesth. 2007;99:68–74.
14. Armstead WM. Cerebral blood flow autoregulation and dysautoregulation. Anesthesiol Clin. 2016;34:465–477.
15. Cole CD, Gottfried ON, Gupta DK, et al. Total intravenous anesthesia: advantages for intracranial surgery. Neurosurgery. 2007;61:369–377; discussion 377–378.
16. Mavrocordatos P, Bissonnette B, Ravussin P. Effects of neck position and head elevation on intracranial pressure in anaesthetized neurosurgical patients: preliminary results. J Neurosurg Anesthesiol. 2000;12:10–14.
17. Li X, von Holst H, Kleiven S. Influence of gravity for optimal head positions in the treatment of head injury patients. Acta Neurochir (Wien). 2011;153:2057–2064.
18. Papadopoulos MC, Saadoun S, Binder DK, et al. Molecular mechanisms of brain tumor edema. Neuroscience. 2004;129:1011–1020.
19. Verbalis JG. Brain volume regulation in response to changes in osmolality. Neuroscience. 2010;168:862–870.
20. Todd MM. Hyperosmolar therapy and the brain: a hundred years of hard-earned lessons. Anesthesiology. 2013;118:777–779.
21. Grande PO, Romner B. Osmotherapy in brain edema: a questionable therapy. J Neurosurg Anesthesiol. 2012;24:407–412.
22. Hernandez-Palazon J, Fuentes-Garcia D, Domenech-Asensi P, et al. A comparison of equivolume, equiosmolar solutions of hypertonic saline and mannitol for brain relaxation during elective supratentorial craniotomy. Br J Neurosurg. 2016;30:70–75.
23. Raghava A, Bidkar PU, Prakash MS, et al. Comparison of equiosmolar concentrations of hypertonic saline and mannitol for intraoperative lax brain in patients undergoing craniotomy. Surg Neurol Int. 2015;6:73.
24. NCCN Clinical Practice Guideline in Oncology (NCCN Guidelines):Central Nervous System Cancers (Version I 2017). Available at: Accessed August 18, 2017.

mannitol; supratentorial brain tumor; brain relaxation.

Supplemental Digital Content

Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.