Effects of positive pressure ventilation on pulmonary physiology

This chapter addresses section F10(i) of the CICM 2017 syllabus document, “Describe the physiological consequences of intermittent positive pressure  ventilation and positive end-expiratory pressure”. The act of forcefully pushing air into somebody's chest cavity has a range of effects, of which some will hopefully be beneficial. Usually, by subjecting somebody to pressurised oxygen torture we expect to achieve some sort of improvement in the pulmonary mechanics or gas exchange. That, after all, is why we subject people to this pressurised gas torture. Specific indications for positive pressure ventilation (and the indications for mechanical ventilation in general, which are subtly different) are discussed elsewhere. This chapter is concerned purely with the effects of positive pressure on the respiratory system. Specifically, generic "positive pressure" effects will be discussed. The effects of PEEP, for example, are slightly different from the effects of positive pressure in general.

Though a fairly important topic, this has only ever emerged once in the history of the CICM Part 1 exam, as a fragment from Question 3 from the second paper of 2014 (“Describe the physiological consequences of Positive End-Expiratory Pressure” they asked). The CICM WCA document (“Ventilation”) also fails to engage with this material. Contrary to this level of official attention, the topic is important. The physiological effects of positive pressure on the lung are also the mechanisms by which mechanical ventilation has its positive effect on the patient. In other words, it seems important to know the functional basis for the thing you’re prescribing. That’s probably a fundamental rule of some sort.

In summary:

Invasive ventilation decreases, and NIV increases anatomical dead space

The anatomical dead space is the long vaguely tubular volume of useless airway which does not participate in gas exchange, and which is therefore “dead”  to the respiratory physiologist. This volume for most people is about 2ml/kg, or 150ml on average, and consists of everything from the lips and nostrils all the way down to the terminal bronchioles.

From this, it follows that if you replace the uppermost parts of that volume with some sort tubular conduit (let’s call it a tube) which has a smaller internal volume, then the total dead space volume should decrease. In short, intubation decreases anatomical dead space.

Kain et al (1969) were able to demonstrate this effect with intubated patients undergoing halothane anaesthesia. When compared to breathing through a mask, the dead space volume decreased by 82mls on average, i.e. it was more than halved. This makes sense. To use the conventional equation for the volume of a cylinder, the 30cm endotracheal tube with an internal diameter of 0.8cm has a total volume of 15ml; to which you might only need to add about 5-10ml of Y-connector.

On the other hand, the use of a large NIV mask increases the anatomical dead space volume. Exactly how much obviously depends on the volume of the mask, and this can vary from 10-20ml with nasal masks to 50ml or so with some of the smaller face masks, and up to litres with the full head helmets. Saatci et al (2004) determined that with a common form of face mask, anatomical dead space increases from 32% to 42% of the tidal volume, or roughly from 150 to 200ml.

The fact that this is not strictly speaking an effect of pressure is not lost on the author, but it is an effect of mechanical ventilation, and is therefore almost always associated with the application of positive airway pressure. Regardless of whether ends up happening invasively or noninvasively, there will be some change to the anatomical dead space, and this needs to be acknowledged.

Positive pressure ventilation increases the gradient for gas transfer

But not by much.

This is a concept mainly related to the theoretical determinants of gas diffusion rather than the practical expectations of the pragmatic intensivist. As such, one could easily omit this from their exam answer and receive no penalty. It has been left out of the summary paragraph at the front of this chapter because it occupies that fringe of interesting but irrelevant material which surrounds real exam-oriented information.

Basically, at atmospheric pressure and normal air gas concentration, one might expect the partial pressure of alveolar oxygen to be about 100mmHg. This, then, is the driving pressure gradient for oxygen diffusion into the lung capillaries. By Fick’s law, that gradient is a relatively important determinant of the rate of diffusion.

Now, consider that by giving somebody positive pressure, you effectively increase the atmospheric pressure, and therefore inside the alveoli the pressure also increases. The partial pressure of oxygen is therefore augmented by this added pressure. The pressure driving diffusion is now 21% × (barometric pressure + PEEP). Of course, this contributes minimally: converted into mmHg, a PEEP of 10 cm H₂O adds 7.4 mmHg of pressure. Under ideal circumstances, if one is ventilating their patient with 100% FiO₂ and a PEEP of 10, the driving pressure of oxygen increases from 713mmHg to 720 mmHg, which not likely to make any difference whatsoever.

Was it pointless to mention this piece of trivia? Perhaps. Let’s say one is under conditions where the ambient atmospheric pressure is for some reason reduced. The relative contribution of that paltry 10 of PEEP might suddenly become relevant. Such a scenario was tested by Schoene et al (1985), who walked thirteen healthy volunteers up Mt. McKinley in Alaska (now known as Denali, its name in the language of the indigenous Koyukon people). The altitude at the research station was 4,400m above sea level, where the atmospheric pressure would be around 440 mmHg (and the room air alveolar O₂ therefore around 57 mmHg, with the measured CO2 of 20). Unfortunately, in Schoene’s subjects the CO₂ decreased as well. The investigators observed an increase in SaO₂ from around 85% to around 88%, which attributed the increase in saturation to increased minute ventilation. Their conclusion was that it might be better to hyperventilate voluntarily at altitude, instead of using a PEEP device.

Of course, this was the 1980s, and up on a mountainside. More recently and under more controlled circumstances, Nespoulet et al (2013) were able to demonstrate that PEEP does in fact improve oxygenation under hypobaric conditions without increasing the minute volume. Again the altitude was similar (4,350m). With a PEEP of 10 cm H₂O, the SpO2 increased by 6-7% on average. Was this because of diffusion? The authors thought so, but their experiment had many limitations. Ultimately, it is possible to conclude that if this mechanism plays some role in the improved oxygenation of patients, that role is surely minor and forgettable. 

Positive pressure ventilation (specifically PEEP) increases the functional residual capacity

The FRC is the volume of air left in the lungs at the end of passive expiration. From this, it stands to reason that any impediment to expiratory air flow should increase this volume of air, by standing in the way of its exit. One such impediment is PEEP. Positive end-expiratory pressure modifies the gradient for normal expiratory gas movement.

Consider: expiration is a totally passive process. It is driven by the elastic recoil of the chest wall and lung tissue. The gradient which drives air flow is therefore between this relatively low intrathoracic pressure and the extrathoracic pressure which is assumed to be zero. Thus, if one increases the extrathoracic pressure, one decreases the gradient for expiratory air flow.

This has been measured empirically. Satoh et al (2012) measured the changes in FRC with intubated patients; from their original paper, the graph below was borrowed without permission and lightly vandalised. The authors measure FRC before induction, after intubation with zero PEEP, at 1 hour with a PEEP of 5, at 2 hours with a PEEP of 5, and then the same with a PEEP of 10.  Obviously, with zero PEEP the FRC was lowest (lower than with spontaneous breathing. Judging from these data, adding a PEEP of around 7.5 maintains the FRC at pre-intubation values. A PEEP of 10 increases the FRC to something slightly higher than the normal spontaneously breathing levels.

Because the FRC is something of an oxygen storage reservoir, the FRC has some implications for the peri-intubation period. Having a nice big FRC full of oxygen allows one extra time to fiddle around with the airway (Sirian et al, 2009). However, in the ICU, the patient is probably already intubated. What, then, is the point of having a higher FRC?

Well. An increased FRC improves alveolar recruitment, thereby increasing the alveolar gas exchange surface and potentially improving ventilation/perfusion matching in those regions of the lung which were previously untouched by gas, by being collapsed. The benefits of having an increased FRC are largely related to the improved and more consistent aeration of alveoli which would have otherwise collapsed at the end of expiration.

Positive pressure ventilation (specifically, PEEP) increases alveolar recruitment

In order to explain this matter, one must resort to the use of the familiar respiratory volume spirogram. The specific concept which needs to be discussed is the closing capacity, i.e. the lung volume at which the small airways begin to close and obstruct. This closing capacity is a volume: it is the minimal lung volume for maintaining open airways. In most people, this volume is below the FRC, but as we age (or as we develop lung disease, or as we lay in an ICU bed for a prolonged period) the closing capacity increases. With age and lung disease the closing capacity ends up being somewhere in the middle of the normal tidal volume range. In short, the FRC decreases to some volume below the volume of the closing capacity.

Having your FRC below your closing capacity gives rise to all sorts of unpleasantness. For one, some alveoli collapse at the end of every expiration, which means during the expiratory phase (two-thirds of the respiratory cycle) they may not be exchanging any gas. That’s wasted time. On top of this, more work is required to re-open them, as compared to just inflating already opened alveoli.

The application of PEEP increases the FRC as already mentioned. With this, one can crank up the PEEP to a level sufficiently high so that the FRC is increased to above the closing capacity, and nothing collapses on expiration. This is clearly going to have some benefits for oxygenation and ventilation.

Positive pressure ventilation improves lung compliance

By increasing the FRC, positive pressure ventilation improves lung compliance. This is thought to be because of the fact that it is usually easier (i.e requiring less pressure) to increase the volume of already inflated alveoli than it is to recruit collapsed alveoli.

Observe the pressure-volume loop at a low PEEP. Let’s say for some reason you’re ventilating the patient at low peep and with volumes below the closing capacity, i.e. there is a large proportion of collapsed terminal bronchioles and alveoli. In this scenario, a large change of pressure is required to recruit the collapsed alveoli. 

Now, increase the PEEP to increase the FRC above closing capacity.

The higher PEEP has increased the FRC to a point where the expiratory volume is now higher than the closing capacity, and the alveoli are all “open” even at the end of expiration. This is the essence of the positive effect of PEEP on compliance.

Now, let’s say you increase the PEEP to some absurd level. 

The lung now gets overdistended. The alveoli are already stretched to capacity and can inflate no more. A large amount of pressure is used to distend them even further, generating a relatively small volume for a larger pressure.

This is all fine and good with abstract pressure/volume loops, but does it work in real life? Apparently, yes.  Suter et al (1978) demonstrated this relationship experimentally. The figure below has been shamelessly plagiarised from their 1978 paper. It demonstrates exactly what the hypothetical pressure-volume loops suggest. The investigators ventilated patients with different levels of PEEP, using a tidal volume of 15ml/kg. In this experiment, lung compliance was best at a PEEP around 6, and then deteriorated (probably because the lung was getting overdistended with such a high PEEP and tidal volume).

 

The theory behind why this happening was initially based on the idea that alveoli – once they are “opened”, i.e. recruited from the collapsed state – will all expand in unison, “isotropically”, when pressure is applied. This isotropic volume expansion is supposed to require less pressure than re-inflating collapsed alveoli. This explanation was first suggested by Glaister et al (1973), whose disembodied monkey lung behaved exactly in this classical fashion. As can be seen from the original diagram, at around 6% of the vital capacity, there is an inflection point. The airway pressure at this point was around 2-7 cm H₂O.

An important thing to mention is that the sigmoid shape of the pressure-volume inspiratory curve and the hump-shaped relationship between PEEP and compliance is something which becomes more relevant in cases of ARDS-like lung pathology, as compared to healthy lungs. The more “normal” your lungs, the more of their pressure-volume curve is described by the steep middle gradient, i.e. the compliance remains good at a wider range of pressures. Suter et al had mainly ARDS-like patients (6 had traumatic lung contusions and 5 had a variety of pneumonias). Matamis et al (1984) were probably the first to publish these sigmoid pressure-volume relationships; their patients were all ARDS patients. Glaisner’s dead monkeys were also ARDS-like, as “excised lungs become somewhat depleted in surfactant and are therefore more prone to closure than when they are in the chest.” Slutsky et al (1980) gave ARDS to dogs using an infusion of oleic acid and found that the closing volumes and pressures increased. “The decrease in compliance at low lung volumes, associated with an unchanged compliance at higher lung volumes, was compatible with the idea that airway closure occurred at higher values of Ptp after oleic acid” they wrote. In summary, the more diseased the lung, the higher the closing pressure and therefore the higher the PEEP requirement.

Positive pressure ventilation may reduce the work of breathing

Or, it may not.

One of the reasons we mechanically ventilate people is to decrease their work of breathing. Obviously, being sedated and paralysed doesn’t count, as those people’s work of breathing is nil. It probably also doesn’t count for normal healthy people at rest; for those guys, the work of breathing accounts for 1-2% of their total body oxygen consumption (Robb, 1997)  or about 2.4 Joules per minute (Mancebo et al, 1995). However, the work of breathing may increase fifty-fold in disease states, and this is where positive pressure ventilation can do something useful. Or, it can do something counterproductive. And virtually always there some combination of the two. Without reproducing the Campbell diagram here, work of breathing and the influences of positive pressure can be described as follows:

Element of WOB	Positive influence	Negative influence
WOB vs. airway resistance	Decreases WOB in inspiration, by increasing the diameter of airways	Increased WOB in expiration (against increased airway pressure)
WOB vs. tissue resistance	Decreased WOB because of improved lung compliance at moderate lung volumes	Increased WOB because of decreased lung compliance at high and low lung volumes

Positive pressure ventilation (even without a bi-level support component) decreases the work of breathing done against airway resistance. There are multiple mechanisms for the overall improvement in the wheezy patient’s condition due to positive pressure ventilation, which are all described elsewhere, but the specific component of work done against airway resistance is because the airways increase in diameter, decreasing their resistance to airflow and decreasing the turbulence of the flow. Barach & Swenson (1939) found that a CPAP level of 7 cm H₂O increased the diameter of small bronchi by 1mm; Lin et al (1995) confirmed that this improves FEV₁.  However, the work of breathing in expiration may increase – the effort of emptying one’s lungs becomes greater when there is increased airway pressure pushing back.

As is discussed above, the improvement in lung compliance at some sort of “happy medium” range of respiratory volumes is something which is expected to decrease the work of breathing. However, most patient lungs don’t behave in quite the same classical manner, and it is unclear where the inflection point or closing volume is for these patients. In short, one is never quite sure whether one is definitely above the closing volume, or below the threshold of alveolar overdistension. As the result, one can never be sure that one’s positive pressure is not actually creating a situation where the lung compliance is reduced, and work of breathing is increased. In the old (2^nd!) edition of The Intensive Care Manual, Oh (1985) maintained that positive pressure ventilation reduces lung compliance, probably because in those days people were ventilated with insane tidal volumes (15ml/kg). In their defence, as Michael Auth has pointed out, those ancient steampunk ventilators never had any automatic tube compliance compensation, meaning the patients never got the full 15ml/kg. 

Positive pressure ventilation may improve the V\Q matching

The supine patient being mechanically ventilated will get atelectasis of the lung bases. This is a well-established fact.

Atelectasis is multifactorial, but most likely the main factor is compression. The bases of the lungs, where the pressure from overlying structures is greatest, will collapse. These are also the regions receiving the greatest amount of blood flow, because of the hydrostatic pressure gradient between the apex and base (even in this 30° head up recumbent position). This blood flow represents shunt: it will not get oxygenated, and will return to mix with the systemic circulation in its unchanged hypoxic state, decreasing the total oxygen content of the arterial blood.

By increasing the FRC above the closing volume, positive pressure ventilation should improve V/Q matching (i.e. diminish shunt). Regions of lung which were previously collapsed for some or all of the respiratory cycle should remain ventilated with PPV. This has been proven numerous times in the literature. Kumar et al (1970) found their patients’ shunt decreased by a mean 9% when subjected to PEEP of 13. In a more recent entry by Spadaro et al (2016) the investigators were able to demonstrate the same phenomenon in patients undergoing laparoscopic surgery (a PEEP of 10 was required).

Positive pressure ventilation may worsen the V\Q matching

By the same mechanism as described above, mechanical ventilation could very easily make oxygenation and shunt worse. It does this by increasing the size of West’s Zone 1. The high pressure applied to the well-ventilated lung puts pressure on the local pulmonary capillaries, increasing the regional pulmonary vascular resistance and creating a situation where the path of least resistance for pulmonary blood flow is through the poorly ventilated lung.  Nieman et al (1988) demonstrated this effect experimentally with 15 cm H2O of PEEP in an animal model. “Microscopically, alveolar capillaries appeared compressed and flattened by PEEP”, they wrote.

Though usually this would refer to the collapsed bases of lung, West’s Zone 1 does not necessarily obey geographic boundaries. One can imagine a situation where a large pneumonia has taken out one whole lung, and though high pressure is used to ventilate the patient all of that pressure ends up distending the good lung while all the blood flow ends up being redistributed into the consolidated lung. There is a lot of literature reporting on this, a representative example is Kim & Heyman (1989) who were able to demonstrate this phenomenon via a radionuclide VQ scan. Their patient, who was on a PEEP of 18 cmH₂O, had reduced perfusion in the ventilated lung, and plenty of perfusion in the non-ventilated consolidated lung.

Minor and hypothetical mechanisms

Of the already irrelevant series of digressions in this chapter, what follows is especially divorced from the pragmatic reality of exam preparation. It is, for the purposes of completeness, important to address these issues. However, if one were to write them into some sort of exam answer, one would probably attract no marks. The reader burrows into this material at their own risk, understanding that their time is being wasted.

Positive pressure increases the total alveolar gas exchange surface

This happens through two main mechanisms:

1. by recruitment of collapsed alveoli
2. by an increase in the total alveolar surface area

The latter assertion stems from the assumption that alveoli are roughly spherical, and from the mathematical relationship between the diameter of a sphere and its surface area. To put it simply, the alveoli distend under pressure and the increase in alveolar diameter results in an increase of the gas exchange surface. There is more membrane to diffuse through; not only that but the capillaries running along the surface of the alveolus get stretched, allowing blood a longer time of exposure to the alveolar gas. Gas exchange improves as a result.

Does this really happen? Obviously, the alveoli must distend, but by how much? It has generally been believed that the increase in lung volume during a breath is something that happens because all the alveoli distend to some reasonably equal degree, i.e. each little alveolus increases its diameter by a couple of microns, and the sum total of all these changes is the net increase in lung volume.

This belief comes from early studies such as Forrest et al (1970), who froze and sectioned guinea pig lungs to determine what happens to alveoli during deep inspiration, and Storey & Staub (1962) who did something similar to cats. At deep inspiration, the alveoli doubled in volume, increased in diameter by 30%, and their surface area increased by 70%. The physiology textbooks of the subsequent decades repeated this assertion, even though the data came from the lungs of small flash-frozen animals, and though the authors cautioned that “ideally the gross mechanical events occurring during the cyclical inflation and deflation of the lungs should be studied during life”.

Unfortunately, when “studied during life” these phenomena are not observed. Carney et al (1999) gazed at the subpleural alveoli of anaesthetised dogs, through microscopes. The photographs to the left are stolen shamelessly from their original paper. The authors compared photos taken at residual volume (maximum expiration) to those at 80% of lung capacity. The alveoli increased in volume only up to about 20% of the TLC, and thereafter remained largely the same size while the total lung volume continued to increase. The authors were forced to conclude that (at least at the range normal lung volumes for mechanical ventilation) the increase in total lung volume must be the consequence of recruitment of previously collapsed alveoli, instead of the result of all the alveoli expanding in isotropic unison. Further research by Gatto et al (2004) confirmed that alveolar diameter does not change overmuch even under very high tidal volumes, and that the change in total lung volume is instead explained “by a combination of mechanisms, including changes in the size of the alveolar duct, alveolar shape changing from a deep cup to a shallow cup without a change in diameter, normal recruitment/derecruitment, and alveolar crumpling and uncrumpling.”

Positive pressure redistributes lung water out of alveolar interstitium

Under virtually all circumstances, the dominant effect of positive pressure ventilation on pulmonary oedema is related to the cardiovascular and compliance effects. However, in addition to these, there is probably a minor effect which results from the redistribution of lung water. To put is simply, the positive alveolar pressure pushes the lung water out of the way, well into the interstitial spaces which do not contribute to gas exchange.

To illustrate, the author is reduced to childish bucket diagrams. Behold, the two compartments of extravascular lung water.

One is the compartment directly abutting the alveolus, which is a reasonably small compartment. This is the alveolar interstitium.  Let us say that it is oedematous for whatever reason.  The extra water in this compartment occupies space and delays the diffusion of gases by increasing the diffusion distance.

Now let us apply some pressure to this compartment.
 

The pressure squeezes the water out of the alveolar interstitium and pushes it into the more compliant peribronchial (and perihilar) interstitium. The result is an improvement of diffusion, convention dictates.

How did we arrive at this theory? Again, by looking at the frozen lungs of dogs. Malo et al (1984) compared lung pre- and post-PEEP. The alveolar septum thickness was greater without PEEP. The authors concluded that PEEP redistributes the excess alveolar water “into the compliant perivascular space, thus eliminating the obstacle to pulmonary O₂ transfer.”

As already mentioned, this mechanism probably plays a minor role in the overall effects of positive pressure ventilation on the oedematous lung. There is probably some minimal redistribution of oedema fluid to the perivascular space, which probably contributes minimally to the improved gas exchange and patient comfort associated with the use of CPAP in cardiogenic pulmonary oedema. Of the literature dealing with the mechanisms of how CPAP works in this condition, there is virtually no mention of this mechanism.