In the context of cardiac anesthesia and surgery, propofol has been found to be a mild vasodilator [33] and to reduce oxygen consumption during hypothermic CPB [34]. Unlike the volatile agents, propofol retains myocardial contractility at clinically relevant concentrations [35] and does not alter the arrhythmogenic myocardial threshold , stabilizing Ca⁺⁺ homeostasis [36]. As a potent scavenger of oxygen free radicals [37], propofol also attenuates ischemia-reperfusion injury, which helps to reduce oxidative stress during the intra- and postoperative phase [38].

Changes in PK behavior during CPB have been demonstrated for almost all anesthetic agents, with the choice and volume of priming fluids and temperature management essential determinants of the distribution, metabolism, and free fractions of these drugs [39]. The PK of propofol during CPB is not fully understood, as results from studies are conflicting. Early studies have shown that the total concentration of propofol is likely to decrease when commencing CPB due to hemodilution and an increase in the free fraction [40] or remain unchanged [41]. Pre-bypass steady-state values are reestablished during rewarming. Bailey and colleagues undertook a PK analysis and were able to quantify this step change of initiation of CPB with an increase of the size of the central compartment from 6 to 15 l and an increase in elimination clearance [42]. There is an offsetting effect of reduced hepatic extraction with graded levels of hypothermia, as hepatic blood flow decreases by almost 20 % after CPB is instituted [43, 44]. Hypothermic CPB (32–34 °C) also seems to alter the pharmacodynamics of propofol, with a higher central nervous system sensitivity during and immediately after bypass than with normothermic off-pump procedures. In a more recent study by Barbosa and colleagues, fairly similar blood concentrations were obtained the majority of time using TCI , despite a higher metabolic clearance during CPB. They found evidence of enhanced sensitivity to propofol using a bispectral index-guided E _max model of maximal drug effect [45]. However, they also demonstrated that PK model-based TCI systems for propofol can safely and effectively be used both during and after CPB. In comparison, volatile anesthetic agents show even larger variations in uptake and elimination when used on CPB [46] and depend heavily on the choice and use of oxygenators [47].

Sufentanil and remifentanil are popular opioid components recently used for TIVA in cardiac patients. Apart from providing analgesia and hemodynamic stability, these modern, potent fentanyl congeners contribute other beneficial physiological effects during cardiac surgery. For example, activation of the delta-opioid receptor can elicit pre- and post-conditioning, contributing direct cardioprotective effects. The role of this mechanism is currently an area of active research.

Since the PK of sufentanil and remifentanil have been well characterized in PK models in the past [12, 48], application of both drugs using TCI has become the obvious choice for most anesthesiologists in countries where the drug label has been extended accordingly. As with propofol, adjustments in dosing with TCI during CPB have to be considered. As a highly lipophilic drug with a rather shallow dose-response curve, sufentanil may require adjustments during CPB based less on changes of clearance and calculated compartments than on higher unbound concentrations [49]. Although there is a 17 % reduction in sufentanil concentration during initiation of CPB when using a constant infusion, this effect is short-lived, and the performance of TCI based on a PK model is little affected during the later stages [50].

Remifentanil , in contrast, shows slightly different PK characteristics when used during CPB . This potent opioid , with a high metabolic clearance and tissue distribution, exhibits a significantly increased volume of distribution with institution of CPB. This increased volume of distribution remains increased even after the end of CPB, as noted by Michelsen and colleagues [51]. The PK of remifentanil during CPB is best described with a two-compartment model instead of the usual three-compartment model description. Again, elimination clearance seems to be reduced proportionally to the level of hypothermia . As metabolic clearance for remifentanil is constant and nearly infinite, the benefit of TCI models lies in the calculated loading and maintenance of the central compartment over time.

The scientific foundation for some current and proposed is controversial. Gathering evidence for these topics faced the same challenges in translating basic research into meaningful clinical effects that are faced by those working in neuroscience and behavioral science. A particularly controversial issue is that of the balance of evidence with regard to the ability of intravenous or volatile anesthetic agents to improve outcome.

The cardiothoracic anesthesia literature of the last decade was dominated by publications advocating the cardioprotective effects of volatile agents over TIVA with propofol , supported by evidence that the volatile agents precondition the myocardium or minimize ischemia-reperfusion injury (IRI) [52]. These results, which are either from preclinical work or studies of secondary outcomes, have recently been questioned in terms of their translational to relevant clinical effects [53–56].

The concept of pharmacological interventions as a protective strategy during cardiopulmonary bypass is based on established experimental findings. It has been shown that repeated episodes of brief myocardial ischemia protect the heart against a subsequent more prolonged ischemic insult [57]. The molecular mechanism of this endogenous protective response to ischemia-reperfusion injury (IRI) triggers various signaling pathways involving reactive oxygen species (ROS) and a process that stabilizes mitochondrial membranes leading to decreased activity of mitochondrial permeability transition pores (mPTP) [58]. From this relatively universal concept, potential pharmacological protective interventions have been suggested and derived [59]. In particular, it seemed to be feasible to demonstrate evidence in the laboratory setting that volatile anesthetics provide protection before (preconditioning) and after (post-conditioning) myocardial ischemia [60, 61]. Based on these, preliminary clinical studies found evidence of reduced biomarkers of myocardial injury when volatile agents were used before, during, and after cardiopulmonary bypass, as compared with TIVA based on propofol [62, 63]. Claims of improved patient outcome in terms of survival or clinical myocardial events remained inconsistent. De Hert who initially reported a significant reduced troponin T release with sevoflurane compared to propofol could not demonstrate this in a follow-up clinical trial [53, 64]. Although the study was not powered for this secondary endpoint, the authors of the latter study showed an apparent advantage in one-year survival in the sevoflurane group against TIVA propofol. Unfortunately, this statistically questionable result was included in a highly quoted meta-analysis and was the pivotal study in it shifting the odds ratio of influence on mortality slightly in favor of halogenated agents [65]. Other authors of another meta-analysis were more cautious in their interpretation of the literature [66]. Nevertheless, the American College of Cardiology Foundation /American Heart Association still recommend a “volatile-based anesthetic ” for CABG surgery procedures in their current guideline dating from 2011 [67].

However, recently, larger clinical studies could not convincingly demonstrate clinical benefits in either myocardial damage or outcome with volatile agents compared to TIVA both in cardiac and noncardiac patient populations of high risk. Lurati Buse et al. studied 385 high-risk noncardiac surgical patients and could not demonstrate a reduction in either cardiac biomarkers or major adverse events at 1 year [55]. This was confirmed in a Scandinavian study in vascular surgery patients [54]. In a randomized multicenter study from Italy in 200 high-risk CABG and valve patients, Landoni et al. reported no impact of the choice of volatile versus TIVA-based techniques on a composite endpoint of death, ICU stay, and 30-day and 1-year mortality [68]. Flier et al. concluded from their study in CABG patients that the clinical relevance of volatile cardioprotection is questionable since significant effects may only be detected in very homogenous groups [69]. A recent proof-of-concept study demonstrated only a reduction of inflammatory markers with cardio-specific sevoflurane exposure, but there was a lack of attenuation of release of markers of myocardial cell damage [70]. Similarly, other strategies of cardioprotection, in particular remote ischemic preconditioning (RIPC) , have obtained negative clinical effects despite proof in principal in animal experiments [71–74]. The claim that propofol itself antagonizes or inhibits a potential organ-protective effect by RIPC is not substantive. In fact, no single study proves this hypothesis. One small study suggested that RIPC in 14 patients, as compared to propofol anesthesia in 19 controls, mildly reduced the area under the receiver-operator-characteristic curve at 72 h for the result of a troponin I assay [75]. This phase I study is however too small to allow a definite conclusion [76]. More importantly, while a recent meta-analysis of 15 randomized trials confirmed that RIPC was cardioprotective, it also showed that the use of volatile anesthetics with RIPC attenuated this cardioprotective effect [77].

We need to bear in mind though that propofol itself, as part of TIVA techniques, is a potent scavenger of ROS and thus may influence the degree of myocardial injury and inflammatory response in a different way [78]. Corcoran et al. could demonstrate a beneficial effect of propofol infusion on neutrophil function, lipid peroxidation, and inflammatory response immediately after cross-clamp release and reperfusion in patients with impaired ventricular function. The data show no difference of reduced biomarkers of myocardial injury in TIVA patients compared to those in the isoflurane group [37]. This may cause clinical interference in confirming isolated volatile effects but more likely provide synergistic protection in its own right [79].

Maybe more importantly, the opioids that are part of modern TIVA techniques are also important factors influencing cardioprotection. Opioids have been shown to confer acute and delayed cardioprotection via opioid receptors, effects similar to ischemic pre- and post-conditioning [80, 81]. Remifentanil , in particular, was studied extensively both experimentally and clinically in terms of its cardioprotective effects. Using a rat model of ischemia-reperfusion injury, Zhang et al. demonstrated that remifentanil preconditioning reduces the size of myocardial infarction after lethal ischemia to a similar extent as ischemic preconditioning [82]. Further studies have accumulated evidence on the complexity of opioid-induced cardioprotection via various receptor subtypes that include effects on the oxidative and nitrosative stress balance in the myocardium [83] with remifentanil preconditioning also providing a second window of protection in a delayed fashion [84]. Clinical studies have in part confirmed these assumptions and also suggest a dose-dependent protective effect of remifentanil on markers of myocardial injury [85–87]. A meta-analysis examining the use of remifentanil versus other opioids showed a reduction in troponin release, time of mechanical ventilation, and length of hospital stay in patients undergoing cardiac surgery [88].

Due to the complex molecular pathways of cardioprotection and the likely influence of genotype and other patient factors on these pathways [71], clinical studies so far have overall provided less than convincing proof that a particular anesthetic agent can consistently be associated with a better outcome after cardiac surgery. As risk prediction for cardiac surgery shifts toward biomarker screening alongside scoring individual patients’ risk factors [89], the likelihood that the presence or absence of one particular intervention within the highly complex perioperative process of cardiac surgery will consistently affect mortality still remains highly speculative.

A more tangible outcome after cardiac surgery is the possible effects of anesthesia on the integrity of brain function and, in particular, its impact on neurocognitive outcome after CPB [90]. Although stroke is still considered a rare complication after CPB, various degrees of cognitive dysfunction and decline are rather common and may last longer compared with noncardiac surgery [91]. Conflicting evidence exists as to whether intravenous or volatile agents are implicated in the development of POCD [92]. There is evidence, however, that volatile agents impair peripheral and cerebral microcirculation which may have an indirect contribution to cognitive decline [93]. Based on experimental research, there is now a possible association between exposure to volatile anesthetic agents and the formation of neurofibrillary tangles and amyloid plaques in patients with Alzheimer disease [94, 95]. In contrast, propofol has been shown to elicit direct neuroprotection by attenuating inflammatory responses during CPB [37], by scavenging hydroxyl radicals formed by brain injury [96], and by reducing the infarct size after experimental ischemia-reperfusion (neuro-apoptosis challenge) in the brain [97].

The practical aspects of TIVA for the cardiac surgical patient include adequate but safe titration during induction, consideration of anesthetic requirements and drug interaction during CPB and rewarming, and the smooth transition from intraoperative opioids to postoperative analgesia. As in other specialties, there is no one recipe that fits all patients but a few important points should be discussed here.

Firstly, the availability of TCI has allowed us to titrate small incremental changes in drug concentration with a much higher margin of safety compared to manual infusions [98]. This applies particularly for the induction process. As with any frail, old, or otherwise vulnerable patient, we would allow the propofol infusion to be titrated in small increments from around 2–2.5 μg/ml target concentration when using the Marsh pharmacokinetic model . A slightly higher initial target concentration of 3.0–4.0 is advised when the Schnider model is used in effect-site control as explained elsewhere in this book (see Chap. 11 on Propofol PK/PD). Dependent on individual choice, the concomitant opioid infusion should have achieved satisfactory levels when endotracheal intubation is performed; for remifentanil TCI, this is in the range of 3.0–4.0 ng/ml.