A prospective randomized trial of the optimal dose of mannitol for intraoperative brain relaxation in patients undergoing craniotomy for supratentorial brain tumor resection.

OBJECTIVE Mannitol is used intraoperatively to induce brain relaxation in patients undergoing supratentorial brain tumor resection. The authors sought to determine the dose of mannitol that provides adequate brain relaxation with the fewest adverse effects. METHODS A total of 124 patients were randomized to receive mannitol at 0.25 g/kg (Group A), 0.5 g/kg (Group B), 1.0 g/kg (Group C), and 1.5 g/kg (Group D). The degree of brain relaxation was classified according to a 4-point scale (1, bulging; 2, firm; 3, adequate; and 4, perfectly relaxed) by neurosurgeons; Classes 3 and 4 were considered to indicate satisfactory brain relaxation. The osmolality gap (OG) and serum electrolytes were measured before and after mannitol administration. RESULTS The brain relaxation score showed an increasing trend in patients receiving higher doses of mannitol (p = 0.005). The incidence of satisfactory brain relaxation was higher in Groups C and D than in Group A (67.7% and 64.5% vs 32.2%, p = 0.011 and 0.022, respectively). The incidence of OG greater than 10 mOsm/kg was also higher in Groups C and D than in Group A (100.0% in both groups vs 77.4%, p = 0.011 for both). The incidence of moderate hyponatremia (125 mmol/L ≤ Na < 130 mmol/L) was significantly higher in Group D than in other groups (38.7% vs 0.0%, 9.7%, and 12.9% in Groups A, B, and C; p < 0.001, p = 0.008, and p = 0.020, respectively). Hyperkalemia (K > 5.0 mmol/L) was observed in 12.9% of patients in Group D only. CONCLUSIONS The higher doses of mannitol provided better brain relaxation but were associated with more adverse effects. Considering the balance between the benefits and risks of mannitol, the authors suggest the use of 1.0 g/kg of intraoperative mannitol for satisfactory brain relaxation with the fewest adverse effects. Clinical trial registration no.: NCT02168075 ( clinicaltrials.gov ).