“The practice of anesthesiology prides itself on having a strong basis in science.”
For most anesthesiologists, volatile anesthetic agents remain a core tool that we use on a daily basis. Real-time tidal gas concentration measurement is one great strength of inhalational anesthesia. However, an end-tidal concentration is just one intermediate point in the complex process between a dialed vaporizer concentration and drug effect. While many of the principles of this process are taught during training, we suggest that achieving the highest standards of patient care in our specialty requires an understanding, and the daily application, of the principles of inhalational pharmacokinetics.
In two papers in this issue of Anesthesiology,1,2 Kretzschmar et al. and Baumgardner et al. present data from animal studies exploring the changes in blood concentrations of sevoflurane and desflurane during uptake and elimination of these agents. This work was conducted at the Hedenstierna laboratory in Uppsala, Sweden.
There are two salient features of these studies. As the authors point out, there is very little published in vivo work on elimination of inhalational anesthetic agents, and yet this phase of anesthesia is important for several reasons. As many texts on inhalational anesthesia point out, in contrast to induction, where we can use “over-pressure” to accelerate onset, we are limited in our ability to manipulate the partial pressure gradient in the same way to accelerate elimination and recovery. Second, the primary focus of these studies was blood content. Most studies of inhalational agents use concentrations in end-expired gases as a questionable surrogate for blood partial pressures. However, it is the partial pressure gradient between inspired gas, blood, and tissues that drives onset and offset of effect.
There are several key findings from these studies. First, they show how, at any given cardiac output, overall lung alveolar ventilation quickly becomes the main determinant of the rate of rise or fall of the partial pressure of the agent in arterial and mixed venous blood, and that this effect persists for at least the first 45 min after commencement or cessation of administration of these agents. The authors further find that after 10 min into the washout phase, volatile agent elimination can be represented by a simple model consisting of just two body compartments: the muscle group and the lungs.
Interestingly, the authors have in the process confirmed in vivo the findings of previous theoretical analyses that posed the question, “What effect does cardiac output have on the rate of elimination of a volatile agent?” These studies predicted that a higher cardiac output slowed the washout or rate of fall in partial pressure of volatile agents in the vessel-rich tissues such as the brain during emergence, in a similar fashion to its slowing of washin during induction.3 While computer models are instructive, their inherent simplifications can deceive us and need to be challenged or validated with real measurements to reassure us of their reliability. Kretzschmar et al. have supported this in their piglet model, where the overall alveolar ventilation/perfusion ratio (V.A/Q.) was manipulated by either doubling, or almost halving from baseline, cardiac output using dobutamine or atrial occlusion, respectively, accompanied by more modest opposite changes in ventilation. They found the rate of decline in venous blood partial pressure is slowed by a lower overall V.A/Q., i.e., a higher cardiac output, and accelerated by lower cardiac output. The authors have gone on to interpret this as in vivo evidence that overall alveolar ventilation rate persists into the postoperative period as the rate-limiting step in the elimination of modern volatile agents from the body.
The implications for the physiology and conduct of anesthesia in humans need to be critically examined. First, under clinical conditions, we do not deliberately modulate global V.A/Q. ratio by manipulation of cardiac output, beyond ensuring adequate fluid and volume resuscitation and blood pressure support to attempt to optimize the patient’s circulation. Rather, the tool we have most readily at hand in everyday practice is the ability to modulate overall alveolar ventilation rate through control of lung minute ventilation. What does this do to the variables measured in this study, where global tissue perfusion remains relatively steady? We may have to rely on computer simulations a little longer to answer this question. In the meantime, the data from Kretzschmar et al. are reassuring that the assumptions we have relied upon and taught for some years are largely correct. These results suggest maintaining minute ventilation during emergence will enhance clearance of inhalation agents from body tissues, and this effect continues into the recovery period. Conversely, factors decreasing minute ventilation such as postoperative opioid-induced respiratory depression may slow elimination of these drugs.
A second reason to question simple extrapolation of these findings to our patients is the assumption that the lungs can be treated as a single uniform compartment with no significant variation in V.A/Q. ratio throughout the lung (i.e., no V.A/Q. scatter). Increased V.A/Q. scatter manifests as widening of end-tidal or alveolar to arterial partial pressure gradients for gases being taken up or eliminated by the lung. This simplification is commonly made in computer simulations of inhalational pharmacokinetics and was not unreasonable for their piglet model where the degree of V.A/Q. scatter appears relatively small, as indicated by the alveolar to arterial gradients measured for carbon dioxide. For this reason, the end-tidal volatile agent partial pressures in the piglets are likely to have been close to the measured arterial partial pressure.
However, these assumptions generally do not hold true in the anesthetized patient. As first shown elegantly almost 40 yr ago by Goran Hedenstierna and colleagues in Stockholm using the multiple inert gas elimination technique, significant scatter in distributions of alveolar ventilation rate and cardiac output is the norm even in healthy human lungs under general anesthesia.4 Consequently, relatively large alveolar to arterial partial pressure gradients for volatile agents persist for the duration of surgery.5 Nevertheless, there is no a priori reason to think that the typical reduction in efficiency of lung gas exchange seen in anesthetized humans should contradict the authors’ finding1,2 that alveolar ventilation acts as the primary brake on volatile agent elimination from the body at any given level of cardiac output.
The characteristics of the main phase of elimination found by the authors were relatively rapid and similar to those seen from muscle alone. That may surprise some. Does this mean that obesity and volatile anesthesia are not such uncomfortable bedfellows as is often assumed? In clinical practice, we are generally a long way from equilibrium with our inhaled volatile agents. In particular, the body fat compartment has relatively low perfusion, thus taking a long time to reach equilibrium. Even with desflurane, the time constant for the fat compartment is 22.5 h.6 The effect of this extremely slow uptake into fat has recently been neatly shown by Weber et al., who, using computer simulation with GasMan (Med Man Simulations, Inc., USA), showed that after 10 h of anesthesia, desflurane partial pressure in fat is only about one third of that in the brain.7 This modeling is supported by clinical data, with Lemmens et al. observing that the effect of increasing body mass index on uptake is minimal,6 while McKay et al. found that duration of volatile agent administration had a much greater effect on recovery times than increasing body mass index.8 These various results support the suggestion of the data from Baumgardner et al. that fat plays a lesser role than generally assumed in onset and offset of activity, at least over short-term anesthetic and postoperative time frames.
The practice of anesthesiology prides itself on having a strong basis in science. This compels us to continue to push the boundaries of what we know about what we do, and to practice based on an understanding of the underlying processes rather than taking a one-size-fits-all cookbook approach to drug administration. Kretzschmar et al. and Baumgardner et al. have added valuable insight into sevoflurane and desflurane pharmacokinetics, as well as hints about models that can help simplify our understanding of their pharmacology. In the process, the authors have reminded us of the importance of seeking a better understanding of the physiology behind our practice, and have shown us that there is much in the world of inhalational anesthesia that is yet to be explored.
We conclude by sadly acknowledging the recent passing of Goran Hedenstierna. During a research career spanning several decades, Professor Hedenstierna led an innovative and wide-ranging program of both clinical and preclinical studies with particular focus on the factors that impact lung function and ventilation/perfusion matching during anesthesia and critical care, and the potential interventions to ameliorate these effects. The global anesthesiology community has lost a distinguished leader, and many of us have lost a great friend and mentor, whose warmth, good humor, and generosity of knowledge will be greatly missed.
Support was provided solely from institutional and/or departmental sources.
Dr. Kennedy has received consulting fees from GE Healthcare (Chicago, Illinois) in the last 36 mo. Dr. Peyton has received research consultancy fees from Getinge/Maquet (Gothenburg, Sweden) in the last 3 yr.