This chapter is most relevant to Section F5(ii) from the 2017 CICM Primary Syllabus, which expects the exam candidates to be able to "understand pulmonary vascular resistance and the factors that affect this". Historically, this has come up in two previous SAQs:

• Question 5 from the second paper of 2016
• Question 4(p.2) from the second paper of 2009

Of these, the college comment to Question 4(p.2) was by far the more informative, as it outlines clearly the sort of answer they were looking for, to the point of even giving a page reference (Wests'). It was used to structure this chapter, and is a good model for what examiner comments should look like. The other question leaned more into what resistance actually is, and how it influences pressure. This is also something worth discussing, probably in its own chapter. 

Beyond the Pulmonary Circulation chapter of Nunn's (Ch. 6, p. 89 of the 8th edition), no other resource brings together all these factors in a way which could be used as an alternative free source of the same information.  If one were to insist on not buying any textbooks, one would be forced to trawl through quite a large bibliography of frequently paywalled articles from the 1960s. 

Factors which influence pulmonary arterial pressure

Though Question 5 from the second paper of 2016 asked for factors which influence pulmonary arterial pressure, judging by the college answer what they really wanted was a discussion of pressure in general hydrodynamic terms. "A structured approach to defining and describing the many factors that influence fluid flow and resistance was required to score well", the examiners said. Poiseuille’s law was brought up. It is therefore somewhat weird that specifically pulmonary pressure was asked about in the question, as this might have misled the trainees and sent them into a pointless discussion of hypoxic pulmonary vasoconstriction and suchlike. What follows, therefore, is an effort to explore the factors which affect the pressure of any fluid travelling any vessel, but with an attempt to flavour the discussion with factors which are uniquely pulmonary.

Thus:

• Pressure is the product of flow and resistance.
• Flow in the pulmonary circulation is equal to flow in the systemic circulation, i.e. it is the cardiac output, and is therefore determined by:
  • Heart rate
  • Afterload (specifically RV afterload)
  • Ventricular stroke volume, which is in turn determined by
    • Preload
    • Cardiac contractility
    • Ventricular compliance
• Resistance in the pulmonary circulation is determined by:
  • Proportions of laminar and turbulent flow
  • For turbulent flow, resistance cannot be determined by standard equations, only to say that it increases non-linearly as flow increases
  • Most flow in healthy pulmonary arteries is laminar
  • For laminar flow, resistance is described by the Hagen-Poiseuille equation:
    
    where:
    • Δp is the pressure difference between the arterial and venous circulation;
    • L is the length of the vessel,
    • μ is the dynamic viscosity of the blood,
    • Q is the volumetric flow rate (cardiac output),
    • R is the radius of the vessels
  • Of these, the easily regulated variable is vessel radius, which is affected by:
    • Blood flow 
    • Lung volume
    • Hypoxic pulmonary vasoconstriction
    • Humoural and hormonal mediators (eg. eicosanoids)
    • Drugs (eg. nitric oxide and sildenafil)

Effect of pulmonary blood flow on pulmonary vascular resistance

One's cardiac output may fluctuate from 3-4 L/min at rest to something like 25L/min with exercise. With these fluctuations in flow, pulmonary arterial pressure remains quite stable (Kovacs et al, 2012).  Because pressure is the product of flow and resistance, this means that pulmonary arterial resistance must vary depending on pulmonary blood flow. In a viva scenario or written SAQ answer, one could potentially skirt around having an in-depth understanding of this subject by reproducing this graph:

This is a relatively famous graph, which one should probably be at least passingly familiar with. It comes from a famous 1965 paper by West & Dollery, and is reproduced in some form or another in virtually all physiology textbooks. The weird vertical resistance scale is the consequence of having to covert the archaic units from the original figure into ones which would be more familiar to the users of a modern-day Swan Ganz catheter (dynes/sec/cm^-5). Out of respect for the authors, the original work is also reproduced here.

 

The pulmonary blood flow in the diagram is lower than would be expected (the scale only goes up to 800ml/min) because the data were collected from an isolated lung belonging to one 26-kg dog, held upright in a plethysmograph box, being perfused by the venous blood of another dog. 

So, how does this happen? Surely, you'd expect the smooth muscle of the pulmonary arteries to play only the most minimal role here. There's just not enough of it. Therefore, some other mechanism must produce this fall is resistance. In fact, there are two such mechanisms: distension and recruitment.

Distension of pulmonary capillaries in response to increasing pressure

Elastic distension of pulmonary vessels occurs in response to increased blood flow. They are sufficiently elastic that they can just blow up like balloons. Sobin et al (1972) got a bunch of cat lungs, perfused them with a silicon polymer at different pressures, and then catalytically hardened the silicon mixture to preserve the lung vasculature just as it was. Slices of the lung were then examined to see how the vessel diameter changed with different pressures. Nothing would say this better than the original microphotograph of the sliced cat lung:

At the bottom of the image, one may see some collapsed cat capillaries from a lung which was perfused with a driving pressure of only 5mm Hg; the thickness of the capillary sheet here is less than 6 μm. Above, fat plump capillaries can be seen, with a diameter in excess of 10 µm, from a lung perfused with a pressure of 20 mmHg. In fact, when the pressure-diameter relationship was plotted, it was found to be relatively linear, at least over a range of physiologically plausible pressures: 

It is impossible and probably irrelevant to speculate what would happen as the perfusing pressure were to increase, but the relationship would surely lose its linearity at high pressures, and beyond that there would be some point at which the alveolar capillaries would lose their integrity in a visually spectacular fashion. More relevant is the low pressure territory. Capillaries with a small diameter due to low pressure would also have a higher resistance to flow, and at a diameter of less than 5 µm would probably be too narrow to accommodate red cells. This is what things probably look like in the apices of the lung: narrowed, functionally useless capillaries, with minimal blood flow through them. As blood flow and pressure increases, these previously narrowed vessels increase in diameter and start to participate again in the pulmonary circulation, i.e. they are recruited.

Recruitment of pulmonary capillaries

The diversion of blood flow into new vascular spaces is an attractive explanation for the decrease in pulmonary vascular resistance with increased flow. Flow increases, ergo formerly collapsed capillaries get some blood into them and consequently the total pulmonary vascular resistance decreases.

Famously, two groups of investigators published articles within two months of each other, each with a similar experiment but different conclusions. Konig et al (1993) injected rabbits with nano-scale particles of colloidal gold, then killed the rabbits, and demonstrated that gold particles were to be found throughout the lung capillaries, i.e none of those capillaries were in any sort of "collapsed" state. At the same time, Conhaim et al (1993) perfused some rat lungs with fluorescent albumin, froze them for sectioning, and found that only 33% of the alveolar capillaries were perfused with the fluorescent marker. 

The difference between these two studies was alveolar pressure. Konig et al had whole rabbits and atmospheric pressure, whereas Conhaim et al used isolated rat lungs which they inflated with 15 cm H₂O in order to turn each lung into one big West's Zone One. The importance of alveolar pressure in this process was confirmed by Godbey et al (1995) who used direct microscopy to observe subpleural capillaries at different perfusion pressures and alveolar pressures.  Wherever the capillary pressure exceeded alveolar pressure, the capillary in question had flow in it (defined by the authors as the presence of red cells). This occurred even in the presence of a physiologically abnormal low flow, i.e. the investigators demonstrated that capillary pressure was the most important factor here.

How much recruitment can you get? Turns out, more than you could possibly ever use. Carlin et al (1991) demonstrated that with increased cardiac output, the diffusing capacity of the lung continues to increase without reaching any sort of plateau, i.e. even with a cardiac output equivalent to 30-35L/min the DL_CO  continued to rise. This means that even at peak exercise you have not found the limits of your capillary recruitment reserve.

Effect of lung volume on pulmonary vascular resistance

In short, the relationship between lung volume and PVR can be summarised into three pointform statements and expressed via a helpfully memorable and highly textbook-represented diagram:

• Pulmonary vascular resistance is lowest at FRC
• At low lung volumes, it increases due to the compression of larger vessels
• At high lung volumes, it increases due to the compression of small vessels

This diagram probably originates from Simmons et al (1961), and though the article itself is a description of a dog experiment, this graph which has propagated so far through the textbooks is a highly gentrified and speculative interpretation, not actually derived from experimental data. An accurate representation of the original animal data from that paper looks like this:

Some variation of this is seen virtually everywhere, and each such graph typically a) lacks axis scale labels and b) has different curve shapes in every textbook. So what's the definitive curve, and is it relevant to know about it? The best one can do is track down a publication with the decency to properly attribute their diagrams, and chase the paper which they reference. In the course of doing this, one generally comes across Thomas et al (1961) which was a study of dog lungs perfused with fresh heparinised dog blood under conditions of static inflation. Their original data is presented below, after a little cleaning with Photoshop.

 The idea that large vessels collapse at smaller volumes and small vessels collapse at large volumes can be attributed to Howell et al (1961), who ingeniously managed to exclude the microvasculature by perfusing dog lungs with kerosene. The nonpolar solvent did not get into the small vessels no matter how high the pressure they used (they went as high as 80 cm H₂O), probably because of surface tension effects. One could therefore measure the resistance to kerosene flow, and be reasonably confident that it was confined to the larger vessels. The authors did just this, and discovered that pressure in large vessels dropped as lung volume increased, whereas pressure in small (dextran-perfused) vessels increased.  Irrespective of how this relationship is represented, it is all-pervasive, and one should be able to reproduce some variant of it when asked about PVR in an exam. Under those conditions, the exact shape of the curve does not matter as much as one's ability to talk through the main events across it. 

Let us walk through those events in some narrative fashion.

• At RV, let's say the lung is maximally deflated. The large vessels, which are usually held open by the effects of alveolar septal stretch and parenchymal traction, find their walls rather less supported than they would be at higher volumes. The weight of the lung also presses on them, decreasing their diameter and changing their cross section, increasing the resistance to flow through its effect on their Reynolds number. Some of the increased resistance is probably also due to the effects of hypoxic pulmonary vasoconstriction. The net effect of this level of lung deflation is to increase the pulmonary vascular resistance, though not by much - the large vessels do not contribute very much to the total pulmonary vascular resistance (40% of the resistance happens at the level of the capillaries).
• At TLC, the lung is maximally inflated. The alveolar septa are stretched, and the capillaries within them are squished between hyperexpanded alveoli. The elastic bands of connective tissue which make up the structural scaffold of the alveolar walls are stretched tight, constraining these capillaries and forcing them into a certain shape. Under these conditions, the resistance of the small vessels increases. Large vessels, on the other hand, are stretched open because their diameter is tied to the diameter of the whole lung. As the whole lung increases in size, so these parenchymally tethered vessels are also stretched open. Theoretically, this should decrease the net pulmonary vascular resistance, but because these large vessels contribute minimally to the total resistance, the overall resistance is still increased.
• At FRC, the pulmonary vascular resistance is at its minimum. The forces which compress the little alveolar wall capillaries and the forces which collapse the larger vessels exert the least influence at this lung volume. 

Though this all sounds very plausible, the reader must be reminded that everything in this description is based on speculation and potentially completely unrelated to in-vivo human lung behaviour. Isolated dog lobe preparations and mathematical models were used to generate these plausible-sounding physiological theories, but nobody at this stage has ever been able to demonstrate any of this stuff in a living human lung, to say nothing of relating it to anything clinically relevant.  

Effect of atelectasis on pulmonary vascular resistance

If low lung volumes are theoretically supposed to increase pulmonary vascular resistance, then logically atelectasis (i.e. complete collapse of lung units) should really increase PVR. Indeed, that is what happens. The decrease in blood flow and the increase in resistance can be objectively demonstrated. Woodson et al (1963) measured an increase in PVR of up to 93% in the atelectatic dog lung. However, the mechanisms for this are not the same as what one might predict from the section above.

Mechanical compression and loss of parenchymal traction don't seem to influence PVR at all in the context of atelectasis; virtually all of the changes in pulmonary haemodynamics are due to the hypoxic pulmonary vasoconstriction. This was demonstrated in an elegant experiment by J.L Benumof (1979). As is often the case on this site, the original diagrams from that paper are presented with some minimal amendments to make clearer the events of the experiment:

Dog lung was collapsed by absorption atelectasis, and the drop in flow was substantial (about 60%). The lung was then reinflated with a nitrogen-CO₂ mixture, and the flow remained almost exactly the same, i.e. with the resolution of mechanical compression there was absolutely no improvement in blood flow. Only when oxygen was introduced into the gas mixture was the flow restored to baseline levels. That's a convenient segue into the discussion of hypoxic pulmonary vasoconstriction.

Hypoxic pulmonary vasoconstriction

This property of pulmonary vessels is one of the main difference between them and their systemic counterparts, which generally tend to dilate in response to hypoxia. This weirdness was explored beautifully by Davis et al (1981), who implanted some hamster pulmonary arteries into the cheek pouch of a hamster and demonstrated that in response to hypoxia, the graft arteries constricted while nearby "normal" cheek arteries relaxed. An excellent recent overview of this phenomenon is offered by Tarry et al (2017, BJA).  This is one of those things which works well when described in a sensor-controller-effector fashion:

• Oxygen sensing by some mechanism, nobody is completely sure what:
  • Direct effect on potassium channels, or maybe
  • Mitochondrial reactive oxygen species production, or perhaps
  • Changes in cellular energy state, or maybe
  • activation of a hitherto undiscovered hypoxia-induceable factor
• Regulation of the response by pulmonary endothelial cells, by means of several intermediate modulators:
  • Nitric oxide, which is counter-regulatory (i.e. it promotes vasodilation)
  • Prostacycline, which also promotes vasodilation
  • Endothelin-1, which is a vasoconstrictor acting via G-protein coupled receptors on vascular smooth muscle
• Effector (vasoconstrictor) response by membrane depolarization after sodium ion influx, leading to an increase in calcium concentration and therefore smooth muscle contraction. 

Some points of note about hypoxic pulmonary vasoconstriction:

• HPV is determined by the total regional oxygenation.  Not only alveolar but also mixed pulmonary arterial oxygen content matters, though the latter matters less. Based on different combinations of alveolar and pulmonary arterial oxygen tensions,  Marshall & Marshall (1988) were able to determine that about a third of the stimulus comes from the oxygen in the pulmonary artery, and about two-thirds comes from the alveolar oxygen. In textbooks this relationship is usually represented by an equation, also developed by the Marshalls:
  
  
  where 
  • PAO₂ is the partial pressure of alveolar oxygen, and
  • PVO₂ is the partial pressure of oxygen in the mixed venous blood
• The specific thing which controls HPV is (probably) oxygen tension, not content. In 1952, Duke & Killick perfused some disembodied cat lungs with a dextran solution at different levels of anaemia, some diluted down to a haemoglobin concentration of less than 10g/L. Provided the tension of dissolved oxygen remained stable, the pulmonary vessels did not care. This is logical, as they also have no input into the total oxygen-carrying capacity of the blood, and therefore it would be pointless for them to constrict in response to anaemia or the presence of weird haemoglobin species.
• HPV is produced by an increase in the resistance of small distal pulmonary arteries.  We are probably talking about vessels approximately 100 μm in diameter.  Staub (1985) describes various experiments by means of which the anatomical site of hypoxic pulmonary vasoconstriction was localised, generally by means of ventilating cat lungs with gas  mixtures of varying oxygen content. Pressure measurements taken at different points in the vascular tree then demonstrated that the main pressure drop was somewhere above the small arterioles (30-50 in diameter), as illustrated by this image from Nagasaka et al (1984):
  
  
  As one can plainly see, other vessels also constrict (even venules) and weirdly capillaries are also observed to diminish in diameter, though this is puzzling as they really have no smooth muscle in them and therefore should probably not vasoconstrict. Various explanations have been offered for this (Interstitial cells? Pericytes? Contractile elements in the alveolar wall?) but none have so far met high standards of scientific rigor.
• Hypoxic pulmonary vasoconstriction is a biphasic process. There is an initial rapid vasoconstriction, and a chronic slower vasoconstriction. In textbooks which mention this, there is usually a diagram of this, and it is usually some variant of this diagram from Talbot et al (2005):
  
  This study included twelve healthy volunteers having their PVR measured indirectly by means of echosonography while breathing a hypoxic mixture (their end tidal PO₂ was 50mmHg).  At the beginning of the hypoxic period, within seconds the vessels had started to constrict, and this process reached a plateau of sorts by about five minutes. After this, a more gradual increase in the resistance takes some hours to develop. At the end, the PVR does not return to baseline immediately, and even with normoxia reestablished the pulmonary arteries are still "in spasm" for many hours.
• Hypoxic pulmonary vasoconstriction is itself affected by many factors, which are listed by Lumb & Slinger (2015):
  • It is more vigorous in neonatal/foetal life, and it may be dampened by ageing
  • It appears to be attenuated by hypothermia
  • It is decreased by iron, and iron infusions can decrease the pulmonary response to hypoxia; in return, desferrioxamine can increase pulmonary hypoxic vasoconstriction.
  • It is reduced in the presence of infection, be it systemic sepsis or localised lobar pneumonia

Regional pulmonary arterial resistance in pneumonia and sepsis

The normal mechanism of hypoxic pulmonary vasoconstriction is somewhat disabled by infection. In the opinion of uneducated laymen, this mechanism has always been thought of as a nitric oxide-mediated thing, as sepsis tends to lead to vasodilation by a variety of NO-related mechanisms. However, this may not be so.  McCormack et al (1993) tested this hypothesis by embolising Pseudomonas-encrusted agar beads into the pulmonary vessels of rats. Once a nice pneumonia had grown, the investigators were able to demonstrate that the infusion of a NO synthase inhibitor (L-NMMA) did not reverse the hypoxic pulmonary vasoconstriction by a statistically significant degree. Other mechanisms are also involved, they concluded.

Metabolicand endocrine influences on pulmonary vascular resistance

Various "humoural factors" can influence the tone of pulmonary vessels. Without digressing extensively on each, these are listed here:

• Catecholamines increase PVR
• Arachidonic acid metabolites (eg. thromboxane A₂) increase PVR
• Histamine (acting on H₁ receptors) usually increases PVR
• Substance P
• Neurokinin A
• Adenosine usually decreases  PVR

Apart from hypoxia and dissolved hormone-like mediators, several other metabolic factors influence  the pulmonay vascular resistance:

• Hypercapnia:  Hyman & Kadowtz (1975) found that hypercapnic ventilation increased the pulmonary vascular resistance of anaesthetised lambs, but not by very much. With a relatively high inspired fraction of CO₂ (12-15%, or about 115 mmHg) there was only a modest increase in pulmonary pressure, from 15 mmHg to about 22.5 mmHg. In humans, the effect is perhaps slightly greater; Kiely et al (1996) found that the PVR  increased from 129 to 171 dyne.cm⁻⁵
• Acidaemia:  a lower pH has the effect of sensitising pulmonary arteries, making them more reactive to hypoxia. Rudolph & Yuan (1966) were able to demonstrate that PVR essentially doubles when comparing hypoxia (FiO₂ of 10%) at a pH of 7.42 vs. hypoxia at a pH of 7.19. They achieved this by infusing their newborn calves with lactic acid, 
• Alkalaemia, in turn, has the opposite effect; hypoxic pulmonary vasoconstriction tends to be suppressed (Loeppky et al, 1992)  
• Hypothermia seems to increase pulmonary pressures, although the data we have in support of this appears to be predominantly from animals - for example this paper from Zayek et al (2000) used seven-day-old piglets. Cooling those piglets down to 32-34° C resulted in an exacerbation of their pulmonary hypertension (which was experimentally induced by an infusion of thromboxane A₂). It is unclear whether one may be able to extrapolate this newborn piglet data to the elderly out-of-hospital cardiac arrest patient.

Control by the autonomic nervous system

Pulmonary arteries have both α₁ and β₂ receptors. They are innervated by both the sympathetic nerve fibres arising from the thoracic spine and by the vagus nerve (M₃ receptors). The density of these receptors favours α₁ neurotransmission, and they seem to be distributed mainly around the larger pulmonary arteries. How much does this system actually contribute to blood flow regulation in the lung?  Surely, funnelling raw adrenaline or acetylcholine into the pulmonary circulation has the effect of modifying pulmonary vascular resistance, but under normal circumstances, the role of the autonomic nervous system in the pulmonary circulation is probably limited. Kummers (2011) reviewed the topic and came to the conclusion that the activation of these receptors has its greatest significance as a trophic stimulus, promoting the hypertrophy of pulmonary vascular smooth muscle and thereby contributing to pulmonary hypertension

Effect of blood viscosity on pulmonary vascular resistance

Certainly, one would not start a conversation about factors which influence pulmonary vascular resistance with a discussion of blood viscosity, as it is probably a fairly minor player. It is also not something we routinely measure. However, it does play some role. Hoffman (2011) reviewed this forgotten factor and was able to scrape together a handful of studies in erythropoietin-treated rats which demonstrated that PVR increased with an increase in haematocrit.

Effect of age on pulmonary vascular resistance

Though pulmonary artery pressure increases with age (Lam et al, 2009), it is likely that pulmonary arterial resistance does not. Lumb & Slinger (2015) also mention that hypoxic pulmonary vasoconstriction is more vigorous in the foetal and neonatal age group.

Drugs which affect pulmonary vascular resistance

One typically finds "drugs" as a category among lists of factors which affect pulmonary vascular resistance. Most of these are probably very familiar to all CICM trainees. In the event that at some future point one needs to generate a list of them, they can probably be presented as a table. These drugs will logically fall into two groups, as pulmonary vascular resistance is a one-dimensional number which can either increase or decrease.

Pulmonary vasodilators and vasoconstrictors

Vasodilators	Vasoconstrictors
Nitric oxide Milrinone Levosimendan Sildenafil Vasopressin Bosantan / ambrisantan Prostacycline and its analogs Calcium channel blockers ACE-inhibitors	Adrenaline Noradrenaline Adenosine