Hypotonic solutions, such as the LR solution, are avoided in neurosurgical patients to minimize cerebral fluid accumulation. In contrast, NS, an isotonic crystalloid, has been widely used in neurosurgery because it is thought to reduce the risk of cerebral edema [7]. However, since NS has equal amounts of sodium and chloride (154 mEq/L), hyperchloremic metabolic acidosis occurs when a large amount of NS is administered because its chloride concentration is higher than the normal plasma chloride concentration (96-106 mEq/L).

Numerous laboratory and clinical studies have reported a dose-dependent association between hyperchloremia and the use of NS [8-10]. Hyperchloremic acidosis is associated with acute kidney injury (AKI) during abdominal surgery [9]. In a large, propensity-matched retrospective study of 22,851 patients who underwent a non-cardiac surgery, postoperative hyperchloremia resulted in acute metabolic acidosis, leading to increased 30-days mortality and length of hospital stay [10]. A large retrospective study on abdominal surgery showed that patients treated with balanced crystalloids had better outcomes, including mortality, postoperative infection, need for renal replacement therapy (RRT), need for transfusions, electrolyte imbalance, and acidosis than those treated with NS [9].

Meanwhile, the adverse outcomes of NS were not observed in a randomized control study of critically ill patients [11,12], non-critically ill patients [13], and postoperative patients who underwent neurosurgery [14]. In a recent meta-analysis, the balanced crystalloid solution was beneficial in significantly reducing postoperative hyperchloremia and metabolic acidosis, but the evidence was insufficient to compare the effects of buffered and non-buffered crystalloids on mortality and organ failure [15].

In contrast, balanced salt solutions (BSSs) replace chloride ions with lactate, acetate, and gluconate, which prevents the occurrence of hyperchloremic metabolic acidosis [16]. A BSS is the most common choice of resuscitation fluid in clinical practice [17]. In patients who underwent craniotomy, the NS group had higher sodium and chloride levels and had more patients with marked acidosis than in the BSS group [18].

However, though LR is a balanced crystalloid solution, it is hypotonic. A decrease of 1 mOsm/L in the plasma osmolality results in an increase of 19 mmHg in the pressure of fluid movements across the BBB, and a 3% decrease in the plasma osmolarity results in cerebral edema with a 3% increase in the brain volume and 30% decrease in the intracranial blood cerebrospinal fluid volume [16,19]. Prehospital resuscitation with LR compared to NS was associated with increased mortality in patients with traumatic brain injuries (TBI) [20]. Therefore, LR is not suitable for neurosurgical patients. Instead, isotonic BSS, excluding hypotonic solutions, such as LR, has emerged as a fluid of choice for patients undergoing neurosurgery [21].

An isotonic balanced solution reduces the incidence of hyperchloremic metabolic acidosis and electrolyte imbalances in patients with brain injury, but the intracranial pressure is not different compared with NS [22]. Although a balanced solution has a clear benefit of reducing hyperchloremic metabolic acidosis, its advantage of reducing morbidity and mortality is not clear and requires evaluation.

High-quality data comparing NS and balanced solutions in perioperative and neurosurgical patients are not yet available. Based on the above evidence, although evidence is still lacking, an isotonic balanced solution is preferred over NS in neurosurgical patients because of the lower risk of metabolic acidosis and renal injury.

Large insoluble molecules in colloid solutions increase the intravascular oncotic pressure. In an animal model of brain injury, oncotic pressure reduction without changing the osmolarity increased cerebral edema [5]. Colloid solutions have commonly been used to decrease cerebral edema and improve hemodynamics during neurosurgery [23].

Traditional intraoperative fluid regimens, which include preoperative dehydration, third space loss, and insensible loss, tend to induce a positive fluid balance that is related to postoperative complications [1].

In the recent decade, several randomized controlled studies have compared restricted fluid therapy with liberal fluid therapy in patients undergoing major abdominal surgeries. Brandstrup et al. [2] showed that patients in the liberal group gained body weight and had more complications than the restrictive group.

After this trial, numerous studies on abdominal surgery showed positive results for restricted fluid therapy, leading to a gradual shift to the trend of using fluid restriction during surgery with the concept of zero-balance. However, in two large observational studies, the zero-balance concept has been concerning due to the possibility of worse outcomes, including AKI associated with excessive restriction [56,57].

Recently, RELIEF trial compared restrictive fluid therapy while maintaining perioperative zero balance with liberal fluid therapy [3]. The results showed that the patients in the restriction group had increased rates of surgical site infection and high risks of AKI.

Based on this recent evidence, worse perioperative outcomes have been observed in patients with both overhydration and excessive fluid restriction. Therefore, fluid optimization is essential for perioperative fluid management. It should also be noted that the amounts of administered volume in the liberal and restricted volume therapies were inconsistent and slightly different for each study [58]. In particular, the postoperative weight gain of the restrictive group in an earlier study by Brandstrup et al. [2] was comparable to the liberal group of the RELIEF study [3]. As such, an excessive restriction can result in worse outcomes, such as AKI.

To achieve the optimal fluid volume status, it is essential to avoid overhydration and excessive restriction and develop individually optimized fluid regimens using objective parameters. These objective parameters should be targeted preoperatively and measured perioperatively.

GDFT, a recently emerging fluid regimen, is a type of fluid administration that optimizes pre-defined targets based on directly measured hemodynamic parameters (Fig. 1), such as the cardiac output, stroke volume (SV), stroke volume variation (SVV), pulse pressure variation (PPV), systolic pressure variation (SPV), pleth variability index (PVI), and other factors [1].

Favorable outcomes and decreased costs have been shown for patients who underwent GDFT during a major abdominal surgery [59-61]. Although the certainty of the evidence was very low, a meta-analysis comparing GDFT and restrictive fluid therapy in major non-cardiac surgeries showed that the mortality was slightly low in the GDFT group, and there were no differences between the two groups in the complication rate and length of hospital stay [1]. Unlike other studies, including this meta-analysis, one study [62] found that the total infused volume was higher in the restrictive group (basal crystalloid infusion ranging from 4 to 10 ml/kg/h) than in the GDFT group. A limitation of this meta-analysis was the lack of a definition of restrictive fluid therapy. GDFT consists of a given basal infusion and repeated boluses of fluids (usually colloids) to achieve a predefined target. The basal infusion rate is particularly important to compare GDFT with other fluid regimens.

In two retrospective studies of patients with SAH, a positive net fluid balance was independently associated with poor outcomes [63,64]. However, as it is difficult to compare restrictive and liberal fluid therapies in neurosurgical patients who must maintain euvolemia, recent studies on GDFT have been conducted. There have been some studies to optimize fluid administration using continuously measured dynamic parameters, such as SVV, PPV, and PVI for patients undergoing neurosurgery.

The SVV is a sensitive predictor of fluid responsiveness before and during brain surgery [65-67]. After the induction of anesthesia and before the start of the surgical procedure, the SVV more sensitively predicted an increase of more than 10% in the SV by LR solution infusion compared to the mean arterial pressure, heart rate, cardiac output, and central venous pressure (CVP) in neurosurgical patients [65]. An SVV of 9.5% was concluded as the optimal threshold (sensitivity: 78.6%, specificity: 93%) for predicting a > 5% increase in the SV after a 100-ml colloid solution infusion [66]. The target of the SVV of GDFT can affect clinical outcomes for supratentorial brain tumor resection [67]. Comparing two GDFT regimens for supratentorial tumor resection (with threshold SVV values set at 10 for the low SVV group and at 18 for the high SVV group), the low SVV group had lower postoperative serum lactate levels, shorter length of ICU stay, and a lower incidence of postoperative neurologic events than the high SVV group [67]. Comparing the GDFT group managed fluid by hemodynamic parameters including the SVV with the control group managed fluid by the therapeutic decision of the attending anesthesiologist, the former had less administered fluids, shorter length of ICU stay, lower ICU costs, and lower lactate levels than the control group [68].

The PPV and PVI have also been reported to be good predictors of fluid reactivity during brain surgery [69-72]. Between the CVP group, which maintained a CVP of 5-10 cmH₂O, and the PPV group, which maintained a PPV below 13%, in patients undergoing a brain tumor surgery, the latter had better postoperative hemodynamic stability and less postoperative fluid requirement [69]. The PPV-guided GDFT during supratentorial tumor resection had a comparable brain relaxation scale, low serum lactate levels, more intraoperative fluids, and higher urine output than the standard care group [70]. In the sitting position for neurosurgery, measuring the PPV and PVI with an ear sensor predicted fluid responsiveness well, but the PVI could not be predicted with a finger sensor. However, the PVI measured with an ear sensor was limited by an unreliable signal in 26% of the patients [71].

A study on children undergoing neurosurgery showed different results. Comparing the PVI, ΔVpeak (respiratory variations in aortic blood flow peak velocity), arterial pressure, CVP, heart rate, inferior vena cava diameter, SPV (including delta up [Δup] and delta down [Δdown]), and PPV in pediatric patients undergoing neurosurgery, the PVI and ΔVpeak predicted the fluid response well, but the PPV and other static and dynamic parameters were reported to be unpredictable [72].

Considering that hemodynamic changes are relatively common in neurosurgery, GDFT, which provides individualized optimal fluid status, is a promising fluid management strategy.