Keywords

Goal-directed fluid therapy is a concept where administration of fluid, vasopressors, and inotropes is targeted to achieve an optimal hemodynamic parameter for better tissue perfusion [1]. In high-risk surgical patients, Shoemaker and colleagues reported a lower rate of postoperative complications, fewer deaths, earlier discharge from the hospital and the intensive care unit, and shorter duration of ventilation when perioperative cardiac index and oxygen delivery were increased to >4.5 L/min/m2 and >600 ml/min/m2, respectively [2]. The extraordinary results had drawn a lot of attention, and the approach has since extrapolated to various scenarios. In this respect, goal-directed fluid therapies have been studied extensively for the treatment of critically ill patients and during major surgery. In contrast, few studies have evaluated goal-directed fluid therapy in neurosurgical patients. In this chapter, we reviewed the clinical utility of goal-directed fluid therapy in patients having neurosurgery and receiving neurocritical care.

Which Goal to Direct Therapy?

An integral part of goal-directed fluid therapy is to establish a “goal” to guide interventions. Ideally, the goal has to be clearly defined and prognostically important and can be measured accurately and noninvasively. Figure 1 shows the common parameters that have been used to gauge tissue perfusion during surgery.

Fig. 1
figure 1

Monitors for tissue perfusion

Full size image

Traditional Goals

Arterial pressure is the primary determinant for driving tissue perfusion and is a standard monitor in contemporary anesthesia [3]. The major drawback of using arterial pressure is that there is no consensus on the minimum arterial pressure required to maintain organ function. In this respect, over 130 definitions on hypotension had been reported in the literature [4]. Nevertheless, multiple cohort studies, systematic reviews, and meta-analyses have shown an increased risk of postoperative adverse events with mean arterial pressure <65 mmHg or systolic arterial pressure <90 mmHg for >10 min [5,6,7,8]. In a randomized controlled trial (RCT), individualized blood pressure management with systolic arterial pressure maintained within 10% of baseline (n = 147) reduced the risk of postoperative organ dysfunction by 27% (95% confidence intervals, CI: 6–44%) compared with standard care (treating hypotension only when systolic arterial pressure was <80 mmHg, n = 145) [9]. Anesthesiologists may also monitor end-organ dysfunction using urine output, acid-base balance, and plasma lactate concentration. These measures however lack temporal resolution, require long turnover time, and tend to worsen only after significant hypoperfusion.

Cardiac Filling Pressures

Alternatively, anesthesiologists were accustomed in measuring cardiac filling pressures, such as central venous pressure and pulmonary artery occlusion pressure, to guide fluid therapy. As a measure of cardiac preload, it is assumed that an optimized filling pressure would be important to improve cardiac output and tissue perfusion. However, the absolute filling pressure is dependent on valvular pathology and ventricular compliance and is therefore a poor predictor for volume status or fluid responsiveness [10, 11].

Cardiac Output

Cardiac output is a surrogate marker of tissue oxygen delivery. Conventional measurement uses the indicator (e.g., cold saline, lithium, or indocyanine green) dilution technique, producing intermittent values, and requires insertion of invasive pulmonary artery catheter [12]. Recent technology development has allowed minimally invasive and beat-to-beat measurements of cardiac output [13, 14]. Currently, four methods are commercially available:

  1. 1.

    Impedance cardiography

Transthoracic electrical bioimpedance (TEB) measures the resistance to a high frequency and low voltage current passing through the thorax. The rhythmic changes of impedance correspond to the variations in aortic blood volume during successive cardiac cycles. Therefore, variations in cardiac output will be reflected by the change in TEB [15]. Clearly, TEB signals are affected by electrical interference and may undermine the utility of the device.

  1. 2.

    Doppler ultrasound

The Doppler technique measures blood flow velocity through the aortic valve and the descending thoracic aorta using suprasternal and esophageal probe, respectively [16]. Cardiac output is therefore the product of flow velocity, reference cross-sectional area, and heart rate.

  1. 3.

    Arterial pressure waveform and pulse contour analysis

It has been long recognized that pulse pressure is directly proportional to stroke volume [17, 18]. Currently, three commercially available monitors have been developed to determine cardiac output based on arterial pulse contour. The LiDCO monitor (LiDCO, London, UK) applies the PulseCO™ algorithm to account for aortic impedance, arterial compliance, and peripheral vascular resistance in order to estimate beat-to-beat cardiac output [19, 20]. LiDCO requires regular calibration using subtherapeutic doses of lithium for dye dilution-derived cardiac output [20]. Similarly, PiCCO monitor (PULSION Medical Systems AG, Munich, Germany) uses a proprietary algorithm to analyze the systolic component of the arterial pulse [21]. Intermittent (transpulmonary) thermodilution-derived cardiac output is required for calibration. The Vigileo-FloTrac system (Edwards Lifesciences, Irvine, CA) measures cardiac output with standard arterial catheter attached special proprietary transducer. In contrast to the other monitors, FloTrac does not require additional calibration [22].

In addition, the arterial pressure waveform monitoring allows anesthesiologists to access fluid responsiveness. In this respect, variations in arterial pressure (5–10 mmHg) during respiration are normal phenomena due to the transmission of intrathoracic pressure. In spontaneously breathing patients, arterial pressure decreases with inspiration and increases with expiration. The reverse occurs in mechanical positive pressure ventilation, where pulmonary blood volume is shifted to the left ventricle during inspiration and therefore increases arterial pressure with an increase in preload. The increase in intrathoracic pressure also decreases right ventricular filling, and will decrease arterial pressure subsequently. In patients with hypovolemia, variation in arterial pressure is exaggerated, and therefore derived indices from the arterial waveform (e.g., stroke volume, systolic pressure, and pulse pressure variation) could be used to guide fluid administration (Fig. 2).

  1. 4.

    Partial rebreathing method

Fig. 2
figure 2

Changes of arterial pressure with mechanical ventilation. Systolic pressure variation is the difference in systolic pressures at end-inspiration and end-expiration. Pulse pressure variation is the ratio between the difference and the mean of maximum and minimum pulse pressure. Red tracing, arterial pressure; blue tracing, central venous pressure; green tracing, airway pressure. Arterial, venous and airway pressure tracings courtesy of Dr. Robert Linton and Dr. Nick Linton (http://www.foxlinton.org/cardiac_output/PCOpages/spv2.html)

Full size image

In this method, brief and step changes in carbon dioxide elimination is compared to the changes in end-tidal carbon dioxide tension [23]. The differential Fick method measures pulmonary blood flow and would only indicate cardiac output if the shunt fraction remains constant during the measurement period. Potentially, the technique could be integrated into anesthetic ventilator providing automated breath-to-breath cardiac output readings [24].

Measures of Cerebral Physiology

It should be noted that all the aforementioned technology measure systematic perfusion, and do not indicate cerebral physiology. Nevertheless, several modalities have been developed to measure cerebral hemodynamics, oxygenation, and electrophysiology (Table 1) [25], majority are designed for monitoring head-injured patients.

  1. 1.

    Intracranial pressure

Table 1 Cerebral monitors
Full size table

Intracranial pressure (ICP) can be measured by inserting a catheter into the ventricles, at the subdural or epidural spaces. In addition, a fiber-optic sensor could be inserted to measure the parenchymal ICP [26]. There are risks associated with ICP monitoring. The reported rate for bleeding and infection ranged from 0.5–2% to 1–5%, respectively [27]. Nevertheless, ICP monitoring has been commonly used in the management of severe head injury [28, 29]. It remains difficult to decide on the ICP threshold that will require treatment. In the Benchmark Evidence from South American Trials: Treatment of Intracranial Pressure (BEST:TRIP) trial, strategies aiming to maintain ICP <20 mmHg in severe traumatic brain-injured patients did not improve 6-month mortality, functional, and cognitive performance [30].

  1. 2.

    Cerebral blood flow

Measurement of global cerebral blood flow (CBF) requires imaging technique (e.g., xenon CT). However, regional CBF could be determined by using bedside transcranial Doppler. This is a measure of flow velocity, and the measurement is operator dependent, requiring appropriate acoustic bone window and specific mounting devices for continuous monitoring. Nonetheless, cerebral blood flow velocity has been used to determine cerebrovascular reactivity of the major cerebral vessels, for adjusting lung ventilation and arterial pressure targets during the management of head-injured patients [31]. In addition, transcranial Doppler can be used to detect vasospasm and hyperemia (Table 1). In carotid endarterectomy, the Doppler signal can be used to detect microemboli during arterial clamp release [32, 33]. However, it is unclear what might be the optimal CBF flow velocity to target.

  1. 3.

    Cerebral oxygenation

Global cerebral oxygenation could be determined by sampling of the venous blood drained to the dominant (90% right sided) jugular bulb [34]. A decrease in jugular venous oxygen saturation < 50% is thought to indicate brain ischemia. However, this measurement lacks spatial resolution. Several companies have since produced noninvasive cerebral oximeter that measures regional cerebral oxygenation using electromagnetic radiation (e.g., near-infrared) [35]. It should be noted that cerebral oxygenation is a relative measurement. Furthermore, the commercially available cerebral oximeters use different algorithms and the readings cannot be directly compared. Others have inserted a parenchymal probe with a Clark electrode to measure tissue oxygen tension. Tissue oxygen tension <20 mmHg is generally considered as critical [36].

  1. 4.

    Cerebral biochemistry

Cerebral hypoxia and ischemia lead to anaerobic metabolism, cellular damage, and release of excitatory amino acids. This will lead to a depletion of glucose store and an increase in lactate, lactate-to-pyruvate ratio, glutamate, and glycerol concentrations. By inserting a microdialysis catheter into the brain, it is possible to measure the concentrations of these metabolites and to gauge the extent of cerebral insult within a small brain region [37]. There are, however, no consensus on the thresholds to intervene. Nevertheless, cerebral microdialysis helps clinicians to understand the pathophysiology associated with brain injury and has been used as a surrogate marker for evaluation of new drugs for neuroprotection [38, 39].

Goal-Directed Algorithms

After establishing the goal of interest, the next step is to design an algorithm so that appropriate treatments can be implemented to achieve these goals. Hemodynamic goals (e.g., arterial pressure, cardiac output, pulse pressure variation) are commonly managed with fluid challenges including a combination of colloid or crystalloid. In a systematic review and meta-analysis of 24 RCTs on goal-directed fluid therapy in patients having major surgery (n = 3861), intraoperative use of colloid was significantly higher in the goal-directed group compared with controls, mean difference (95% CI): 467 (331–603) ml [40]. In addition, the administration of vasopressors or inotropes is getting popular to achieve these goals (Fig. 3).

Fig. 3
figure 3

Typical algorithm of goal-directed therapy using systolic pressure variation (SPV) and pulse pressure variation (PPV) in patients receiving general anesthesia with mechanical ventilation

Full size image

In the management of neurocritical care patients, other measures, such as supplemental oxygen, hyperventilation, hypothermia, pentobarbital coma, osmotic therapy, and anticonvulsant, are used to achieve the goals (Fig. 4) [41].

Fig. 4
figure 4

Treatment algorithm (adopted from the Seattle International Severe Traumatic Brain Injury Consensus Conference) for patients with severe brain injury guided by intracranial pressure monitoring [41]. Patients who did not respond will require higher tier of treatments. MAP mean arterial pressure, CPP cerebral perfusion pressure, SpO2 arterial oxygen saturation, Hb hemoglobin, PaCO2 arterial partial pressure of carbon dioxide, ICP intracranial pressure, EVD external ventricular drain, EEG electroencephalogram, [Na] plasma sodium concentration, CT computed tomography

Full size image

Outcomes of Goal-Directed Therapy

Targeting Hemodynamic Variables

Using the hemodynamic targets, >100 studies have evaluated the effectiveness of goal-directed fluid therapy to improve outcomes after surgery [42]. There were also >20 systematic reviews and meta-analyses summarizing these results. Using different combinations of studies on various outcomes, these analyses showed that goal-directed fluid therapy reduced rates of postoperative ileus [43,44,45,46], sepsis or infection [47,48,49], postoperative cardiovascular complications [50, 51], renal failure [48], or any complications [46, 52, 53]. Goal-directed fluid therapy also reduced hospital stay and duration of intensive care unit admission [45, 54,55,56,57], and there was a decrease in hospital or 30-day mortality, compared with controls in noncardiac surgery [58,59,60,61]. In cardiac surgery, overall complication rate and hospital stay were reduced with goal-directed fluid therapy [62]. Other meta-analyses, however, demonstrated no difference between groups [48, 53, 63,64,65,66,67,68,69,70]. Only few trials had studied goal-directed fluid therapy in neurosurgery [71,72,73,74]. Three trials had studied craniotomy [71,72,73] and one on spine surgery [74]. Only two trials reported postoperative outcomes [71, 72]. With limited sample size (total 208 patients), goal-directed fluid therapy reduced a composite of complications (sepsis, stroke, renal impairment, and all-cause mortality), 17% vs 35%, odds ratio (95% CI): 0.38 (0.20–0.73), p = 0.004, I2 = 0.0%.

Nevertheless, in an attempt to pool all 110 trials, Kaufmann and co-workers report large amount of heterogeneity, and it was not possible to perform meta-analysis. Furthermore, the results were sensitive to the studies included, sample size of individual trial [median (interquartile range) size = 40 [30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64] patients per group], monitors or targets chosen, and analytical methods used. Clearly, a large RCT is required to resolve the controversy whether target-guided therapy will improve postoperative outcomes [75, 76]. Two trials are currently ongoing. The FLuid Optimisation in Emergency LAparotomy (FLO-ELA) trial will randomize 7646 patients, ≥50 years to have anesthesia guided by stroke volume variation or control (ISRCTN14729158). Similarly, the OPtimisation of Peri-operaTive Cardiovascular Management to Improve Surgical outcomE II (OPTIMISE II) trial recruits 2502 patients having elective gastrointestinal surgery [77]. OPTIMISE II trial compares 30-day infection in patients receiving fluid and low-dose inotropes (dobutamine or dopexamine) to achieve the targeted stroke volume variation or usual care. The results will inform the role of hemodynamic-guided target therapy on postoperative outcomes.

Targeting Cerebral Variables

Electroencephalogram (EEG) is the commonest cerebral monitor during surgery. EEG-guided anesthesia is thought to be useful in avoiding excessive anesthetic administration and may improve postoperative outcomes. Several trials have evaluated the effect of EEG monitoring in reducing postoperative deaths. The Perioperative Quality Initiative (POQI)-6 conference gathered a number of international, multidisciplinary experts to review the literature on the clinical utility of EEG [78]. In nine trials (ten publications [79,80,81,82,83,84,85,86,87,88,89], n = 8512) that examined postoperative mortality, EEG-guided anesthesia did not reduce all-cause death rate, 14.1% vs 15.1%, relative risk (95%CI): 0.95 (0.80–1.12), P = 0.528, I2 = 38.4%. Since this systematic review, another large RCT was published. The Balanced Anesthesia trial randomized 6644 patients to receive deep or light anesthesia based on bispectral index (BIS) EEG monitoring. The death rate at 1 year after surgery in the deep anesthesia group (7.2%) was not different from the light anesthesia group (6.4%), p = 0.223. Figure 5 shows the updated meta-analysis. In 15,156 patients, the pooled relative risk (95%CI) was 0.94 (0.82–1.07), p = 0.339, I2 = 32.79%.

Fig. 5
figure 5

Forest plots of trials comparing the risk for long-term mortality after surgery in patients receiving EEG-guided or routine care anesthesia. *Patients in the control group received deeper anesthesia; CI, confidence interval; EEG, electroencephalogram; ENGAGES, Electroencephalography Guidance of Anesthesia to Alleviate Geriatric Syndromes; STRIDE, A Strategy to Reduce the Incidence of Postoperative Delirium in Elderly Patients

Full size image

Alternatively, anesthesiologists have used cerebral oximetry to guide anesthetic administration. In a systematic review and meta-analysis of 15 trials (n = 1822), there was no convincing evidence that monitoring improved outcomes, primarily related to the lack of events [35, 90]. Nevertheless, cerebral oximetry monitoring might improve cognitive performance at 1 week after surgery and reduce intensive care unit stay by 5.5–6.9 h (n = 379) [90].

The effect of goal-directed fluid therapy for the management of patients requiring neurocritical care remains less well defined [25]. Majority of the current guidelines recommend multimodality monitoring including ICP, cerebral autoregulation, oximetry, and intracerebral microdialysis [28, 29, 91, 92]. Nevertheless, the quality of evidence underlying these recommendations was low. Majority studies reported association and did not demonstrate causal relationship between goal-directed therapy and outcomes in neurocritical care patients. In particular, it is unclear what would be the minimum monitoring required and the targets that should be adopted (Table 1). Further outcome trials would be required before goal-directed fluid therapy could be widely adopted.

Conclusions

There are numerous barriers in implementing goal-directed fluid therapy during anesthesia and caring for critically ill patients. Further trials should define the goals to target, and feasibility in implementing protocol. Large RCTs are required to define the benefit and risk ratio in adopting the goal-directed fluid therapy in specific patient populations.