As far as fluids are concerned current practice throughout the world is still influenced by the ground breaking paper from Shires et al from Dallas published in 1961 in the Annals of Surgery [10]. The investigators used a triple indicator dilution technique which suggested that there was a profound loss of functional extracellular fluid during major surgery, later nicknamed “third space”, which could amount during surgery to up to 15 ml.kg⁻¹ body weight hr⁻¹. Given that some patients might therefore need 1 L or more per hour of extracellular fluid equivalent replacement they proposed the introduction of a lactated version of Ringer’s solution, a balanced salt solution which contains a more physiological concentration of chloride in contradistinction to so-called “normal saline”, 0.9 % NaCl. It literally opened the floodgates as anaesthetists could point to excessive “third space loss” as the cause of intra-operative hypotension in major surgery and were given carte blanche to use excessive amounts of fluid such as lactated Ringer’s (Hartmann’s) in trying to reverse the fall in blood pressure.

It is clear that the original authors very quickly became concerned that excessive fluids were now been given rather than the fluid restrictive practice which had existed prior to the paper of 1961. Moore and Shires in 1967 stated [11] “the objective of (perioperative) care is restoration to normal physiology and normal function of organs, with a normal blood volume and functional body water and electrolytes. This can never be achieved by inundation (with fluids)”. It is clear that giving massive volumes of balanced salt solutions will not achieve anything resembling normal physiology! The dangers of excessive fluids and volume overload with high CVP have recently been re-iterated by Marik. e.g. pulmonary oedema and increased extra-vascular lung water, impaired oxygenation, altered pulmonary and chest wall mechanics, increased work of breathing, myocardial oedema etc. The list is long [12].

Interestingly, recent papers and reviews have confirmed that “there is no convincing evidence supporting the existence of the non-anatomical third space loss neither in haemorrhagic shock nor in surgery of any kind” [13–15]. This highlights another problem. Third space loss does exist in conditions such as septic shock, anaphylaxis and burn injury and it is important to distinguish fluid and flow requirements in elective surgery in contradistinction to haemodynamic changes in the intensive care unit where sepsis may well be a serious problem. Reviews of fluid management often mix up these two completely different scenarios and thus cause unnecessary confusion.

It is thus disappointing to see a large scale prospective trial of restrictive versus liberal fluids during surgery, the Relief study, which will be running until November 2018 where large volumes of fluid are given in the liberal group. Intervention patients may receive 4 days Na⁺ requirement in a 2-h procedure! In 2016 patients are still receiving too much fluid and Na⁺ in the perioperative period with implications for poor outcome (see later).

It is easy to restore MAP to “normal” using vasoactive agents such as metaraminol, phenylephrine or even noradrenaline [16]. However, MAP may then be maintained at the expense of cardiac output as shown in Fig. 13.1 which demonstrates that reliance on MAP alone may be associated with profound falls in cardiac output (CO) and oxygen delivery (DO₂) and subsequent build-up of oxygen debt (vide infra). It is crucial to know the cause of the fall in MAP before using vasoactive compounds. This may be a real problem when the flow monitor does not accurately track changes in flow with changes in vasomotor tone [17]. This may lead in some cases to a fall sense of security.

Most anaesthetists opine that the fall in MAP following induction is mainly due to peripheral vasodilation with a fall in systemic vascular resistance (SVR) but little change in CO and oxygen delivery DO₂ and incorrectly assume that a simple vasoconstrictor such as metaraminol will both correct MAP whilst maintaining flow (see above). The first measurement of CO during surgery using the dye dilution technique was carried out in 1960 and showed that the CO dropped by 20 % after induction [18]. However, it was only following the advent of continuous CO monitoring technology, particularly using analysis of the radial artery waveform (see later) that it was possible to follow the rapid changes of pressure and flow following induction of anaesthesia. Kamenik and Petrun, using the LiDCO rapid (LiDCO, UK), were able to demonstrate that the fall in MAP following BIS (Medtronic, USA) targeted induction using either propofol or etomidate was mainly due to a fall in CO and not due to a fall in SVR [19]. The fall may be as much as 40 % in elderly, high-risk patients. Knowledge of this potentially dramatic fall is essential to allow appropriate intraoperative fluid and vasoactive drug management of these patients.

To understand what happens to the circulation during anaesthesia induction we need to go back more than 50 years to a series of papers by Professor Sharpey-Schafer at St Thomas’s Hospital Medical School in London [20–22]. He studied venous tone in patients and volunteers who were healthy, anaemic, hypertensive, following Valsalva manoeuvre, ganglion blocking drugs and blood donation. His comment that “it is clear that the usual clinical practice of following such cases (hypotension with ganglion blocking drugs, then being introduced for the treatment of hypertension [23]) by the measurement of arterial pressure by the antiquated cuff method can be of limited value”. He appreciated that the changes in MAP caused by ganglion blocking drugs especially when the patient was raised to the upright position were occurring too quickly to be picked up by conventional methods of measuring MAP in use at the time, such as Korotkoff sounds and the mercury manometer. Interestingly, hypotensive anaesthesia had been introduced into clinical practice by Enderby and colleagues at East Grinstead, Sussex, UK to allow more accurate plastic surgery to be carried out on airmen who had received facial burns during the Second World War. Enderby also noted the restrictions imposed on accurately measuring MAP during these cases and introduced the mechanical von Recklinghausen and Boulitte oscillotonometers into clinical anaesthetic practice in the UK. These he found more accurate at low pressures than the conventional blood pressure cuff [24]. This also allowed remote measurement of blood pressure and removed the necessity of placing the stethoscope over the brachial artery. It is noteworthy that modern automatic oscillo(tono)meters are now the most common way of measuring blood pressure during anaesthesia. Sharpey-Schafer’s conclusion was that “it appears that the veins and not the arteries are the major factor in causing hypotension with ganglion blocking drugs when assuming the upright position”. Is the same true with anaesthetic drugs? Isolated hepatic portal veins and aorta taken from the rat were used to investigate the direct action of propofol. It caused a dose related decrease of tone in both types of blood vessel but concentrations required to produce this effect on veins were significantly lower than those required to produce similar changes in the isolated artery preparation. The authors concluded that this direct action may contribute towards the hypotensive effects of propofol [25]. Venodilation with propofol resulted in an increase in venous capacitance in dogs and there was no change in SVR to account for the change in MAP. They concluded that anaesthesia with propofol is accompanied by decreased CO secondary to venodilation and reduction in preload. Cardiac output and MAP are preserved as long as preload is maintained [26]. It was also noted by Sharpey-Schafer that underlying venous tone is markedly increased in patients with heart failure and anaemia, particularly pertinent to the high-risk and elderly patient population. It is therefore reasonable to conclude that these patients are more susceptible to the venodilation induced by propofol and probably other anaesthetics.

Venodilation is thus a key feature of anaesthesia which has been very little recognised [27]. Following venodilation the fall in MAP is due not only to falling preload, stroke volume (SV) and CO with a consequent reduction in DO₂ but also due to a shift of volume out of the arterial tree into the dilated venous compartment [27]. The question is whether we should maintain venous capacitance with fluids (as in the dog study above) or by administration of a venoconstrictor. Our experience with the venoconstrictor, phenylephrine, the effect monitored by the LiDCOrapid (see later) commenced in low-dose pre-induction can maintain venous tone, cardiac output and MAP without unnecessary fluid administration to maintain venous capacitance [28].

Crucially important is the fact that measuring CO with whatever technology after induction of anaesthesia may greatly underestimate the true resting CO and therefore any intervention which is supposedly designed to “optimise stroke volume” (whatever that means) and maintain DO₂ will be flawed and may result in unnecessary excess fluid and Na⁺ administration to the detriment of outcome in high-risk patients [12].

Although flow monitoring was advocated in high-risk patients by NICE it was not clear what protocol would or should be used by anaesthetists. The principle of goal directed therapy (GDT), see later, had been introduced to try and achieve population-based DO₂ targets and the figure of 600 ml of oxygen delivery per m⁻² Body Surface Area (BSA) is often quoted. However, although this target is usually used in the immediate postoperative period to reduce repay oxygen debt and reduce complications [37] it is less frequently used during high-risk surgery [38] as will be discussed in the next section.

The original Doppler studies suggested that flow time, corrected for heart rate, (Ftc) maintained at about 350 msec seemed to indicate that patients were optimally filled [31]. This important parameter was discarded following the introduction of SV optimisation or maximisation strategies (see Fig. 13.2). This was based on the opinion that poor outcome was a result of inadequate tissue perfusion and giving additional fluids to improve perfusion would result in a better outcome. Thus if the patient was fluid responsive it was assumed that they needed fluids. Thus 200–300 ml. boluses of colloid (3–4 ml.kg⁻¹) are given over a period of less than 5 min and repeated until additional boluses no longer produced a greater than 10 % increase in stroke volume. (A 10 % threshold is chosen as a 300 ml. fluid challenge will produce haemodilution and around a 5 % reduction in oxygen carrying capacity and this would need to be offset by at least a 5 % increase in SV for the fluid challenge to increase DO₂). It is clear that just because the patient may respond to fluids with an increase in SV does not necessarily mean that they need fluids (see also next section). In addition, since the Doppler is placed post induction none of the studies appreciated the extent of the fall in CO and DO₂ post induction.

Another parameter available from monitors showing the dynamic pressure waveform from an arterial line is pulse pressure variation (PPV) [39]. Fluid depleted patients show high values of PPV and this can be used as an indication to give fluids. In addition, if flow is calculated from the pressure waveform (see above) stroke volume variation (SVV) can also be used as an indication of fluid responsiveness. These parameters are only available following intubation and institution of intermittent positive pressure ventilation (IPPV) using a tidal volume of at least 8 ml.kg⁻¹ in a patient with a closed chest and in sinus rhythm. A SVV of between 5 and 10 % suggested adequate filling whereas a figure above 13 % suggested fluid depletion [39, 40]. As will be seen later excessive quantities of unnecessary fluid and Na⁺ have deleterious implications for both the peripheral circulation and the cardiovascular system.

The clinical implications and physiological importance of the endothelial glycocalyx layer (EGL) have been pointed out in three recent reviews and have altered our understanding and management of fluids during surgery [42–44]. The identification of the EGL, a fragile, sieve like barrier between the plasma and the endothelial cells, requires a fundamental re-evaluation of fluid transfer into and out of the interstitial space from arteriolar to venular end as proposed by the Starling equation. It now appears that gross fluid transfer is much less than previously suggested. Fluid crossing the EGL and into the interstitium returns to the circulation predominantly by the lymphatics. The EGL is the true effective barrier between the circulation and the interstitial fluid. As it is fragile it is easily damaged by ischaemia, sepsis, burn injury, anaphylaxis and pertinent to this discussion, hypervolaemia. Excess fluids and Na⁺ to replace non-existent “third space loss” may not only damage the EGL directly by compression and distortion but also lead to an increase in atrial natriuretic peptide (ANP) which itself damages the layer. Albumin appears to be protective and sits within the sieve like membrane partially preventing excess fluid loss. Excess fluid administration in the perioperative period leads to a fall in albumin which again damages the EGL. The implication is that “third space loss” may actually be precipitated by excess fluids rather than they being a replacement for third space loss. Maintaining pre-induction values of CO and DO₂ (e.g. by using colloids/blood to replace obvious losses to maintain circulating volume) rather than trying to maximise them to nominal population based targets may be the more appropriate strategy. This will also help to minimise unnecessary intraoperative fluids and Na⁺ which may be deleterious in the high risk patient.

In experiments in the 1950s Guyton showed that the need of the body for oxygen is the real regulator of cardiac output [45]. In other words, CO is determined by the metabolic needs of the tissues and not the other way round. In general, tissues (except skin and kidney) regulate their own flow according to their metabolic demands. Summation of these tissue flows determines venous return and venous tone determines how much of that venous return actually gets back to the heart as preload. Venous tone is decreased by anaesthesia and small doses of venoconstrictors prevent or substantially reduce the venodilation (see earlier). The heart plays a permissive role and does not regulate its own output, so it is not appropriate to use a SV optimisation or maximisation strategy or try to achieve a perioperative increase in global blood flow as an intervention strategy in the elective patient. It seems much more appropriate to just maintain pre-induction CO and DO₂.