What are the three factors that affect the resistance of blood flow through a blood vessel and how does each factor affect resistance?

This chapter is vaguely relevant to Section G4(vi)  of the 2017 CICM Primary Syllabus, which asks the exam candidate to "describe total peripheral vascular resistance and the factors that affect it". Unlike some of the other syllabus items, this one is finite. You can almost see the sides of it. It has only appeared once in the college papers, in Question 7 from the first paper of 2020, where specifically the physiological control of systemic vascular resistance was asked for.

In summary:

• Vascular resistance is defined using Ohm's Law, where R = ΔP/Q
• Vascular conductance is the reciprocal of resistance, C = Q/ΔP
• Peripheral (or systemic, or total) vascular resistance is the resistance (pressure drop) generated in blood flowing through the whole arterial circulation.
  • Normally the pressure gradient is constant, and the flow is regulated by changes in vascular resistance.
  • This can be measured in Hybrid resistance units (or Wood units), or in dynes.s.cm-5 
• Arterioles (200-100 μm in diameter) are the main site where the resistance develops
• It can be estimated us
• Factors which affect peripheral vascular resistance are the parameters of the Hagen-Poiseuille equation, which is R = (8 l η) / πr4, where: 
  • l = length of the vessel
  • η = viscosity of the blood
  • r = radius of the vessel
• Length of the vascular tree is largely constant, and can be safely ignored
• Viscosity of the blood is affected by:
  • Haematocrit
  • Lipid and protein content of blood
  • Temperature
• Radius of the arterioles is the most important factor, affecting vascular resistance, and it is regulated by systemic and local factors: 
  • Systemic factors include:
    • Arterial baroreflex control  (increased BP leads to a decrease in SVR
    • Peripheral and central chemoreceptors (hypoxia leads to increased SVR)
    • Pulmonary baroreceptors (hypoxia leads to increased SVR)
    • Hormones (eg. vasopressin and angiotensin)
    • Temperature (hypothermia leads to increased SVR)
  • Local/regional factors include:
    • Intrinsic myogenic regulation (in response to stretch)
    • Metabolic regulation (in response to increased tissue demand)
    • Flow- or shear-associated regulation (in response to increased local flow)
    • Conducted vasomotor responses from neighbouring vascular sites
    • Local cooling (which leads to vasoconstriction first, and then to vasodilation again)
    • Immunological modulation by inflammatory mediators

For every tiny aspect of this topic there are multiple highly respected sources clamouring for your attention, and the real skill of the reviewer is to filter out the ones which are both easy to read and sufficiently comprehensive. Clifford (2011) is probably the best free resource for local control of vascular tone. 

Definition of systemic (peripheral? Total?) vascular resistance

Systemic vascular resistance is so named to make it distinct from pulmonary vascular resistance, which is a completely different animal. On occasion, one may also find the textbooks making references to "peripheral vascular resistance", which implies that the vessels of interest lie in peripheral body structures (which, to be fair, most of them do). Alternatively, some authors prefer the term "total vascular resistance", as this term accurately captures the nature of the peripheral circulation, as a massive array of tiny individual resistors mounted in parallel. 

It appears that the concept of vascular resistance is borrowed from hydraulics, which in turn borrowed it from electrical engineering (Ohm's Law). This law states that:

R = ΔV/I

where 

• R is the resistance,
• ΔV is the electrical potential difference, and
• I is the current

Electrical and fluid conductors appear to exhibit similar behaviour. Thus, vascular resistance can be described as 

R = ΔP/Q

where 

• R is the resistance,
• ΔVP is the difference in pressure along the circulation, and
• Q is the blood flow rate

This electrical analogy is probably the most scientifically solid definition you are going to get. Or at least most of the published authors seem to have arrived at this conclusion. Pappano & Weir (the official CICM text)  and Skimming (1997) also define it in this way. As with everything in medicine (or science broadly speaking) even this relatively innocent idea has a toxic comment culture. Posting to an obscure fanzine,  Rogers (2014) opined that the SVR, as it is presently defined (MAP-CVP/CO) is just some sort of mathematical trickery, "is meaningless in clinical practice ...and has no place in the treatment of patients." 

Vascular resistance versus vascular conductance

It is not clear why we settled on resistance specifically, as flow is usually what we are interested in (that's what delivers the blood, after all), and so theoretically we should be all obsessed over vascular conductance instead. Conductance is the reciprocal of resistance:

C = Q/ΔP

As pointed out by Joyce et al (2019), the decision to pick resistance appears to have been completely arbitrary, and probably based more on an idealised model of the circulatory system than empirical biology. Resistance is a suitable index for situations where flow is constant, to describe the changes in pressure gradient which result from changes in vessel diameter. It' is the index you would choose if you were modelling the circulation for the classroom. That's not really what happens in living systems. Instead, flow changes according to changes in vascular diameter, and the systemic pressure remains relatively unchanged. Conductance might actually be better to describe this state of affairs. It is also a better metric to describe situations where flow ceases (at or below the critical closing pressure of blood vessels), where conductance drops to zero (a value which can be mathematically manipulated and has meaning physically), whereas under the same circumstances resistance increases to infinity.

Apart from the inelegance of using infinity in your maths,  Lautt (1989) observed that the use of resistance has led to some misleading interpretations of experimental data. For instance, when Yamaguchi and Garceau (1980) tickled hepatic sympathetic nerves, they were able to demonstrate a markedly decreased conductance (it dropped exponentially with increasing sympathetic stimulation), but the resistance barely budged at all, mainly because the change in pressure was much lower in magnitude than the change in flow. Did vasoconstriction not occur? Of course it did. But, using vascular resistance as your instrument, you would underestimate the effect. Meanwhile, the liver is only getting 75% of its original blood flow. 

These concepts need to be internalised, processed, and put aside for the CICM First Part Exam. There is no risk of meeting an examiner who would insist on a detailed discussion of these concepts. At worst, in some nightmare scenario where a viva station on peripheral vascular resistance occurs, the trainee would probably be expected to reproduce the graph below, and state that in the normal human organism the pressure gradient is constant, and it is the flow that is regulated by changes in vascular tone.

Site of peripheral vascular resistance

Most commonly, when textbooks discuss peripheral resistance, they make some mention of where that resistance occurs, and typically a diagram of blood pressure over vessel calibre is trotted out, such as the one below.

They are usually paraphrasing Davis et al (1986), whose original diagram is presented here as a tribute. The authors measured these experimental data from hamster cheek pouches (lovingly sketched by the investigators) and found that most of the pressure drop occurs in arterioles around 200-100 μm in diameter.  Earlier, Zweifach (1974) experimented with a cat mesentery and determined that in the microcirculation, there is another cliff where a drop in pressure occurred at the precapillary arterioles, vessels with a diameter of 25-15 μm.

This leads to the next point in discussing sites of resistance. Though the CICM examiners remarked that their question "did not require a discussion of individual organs", when one discusses the sites of peripheral vascular resistance, it is important to mention that there are organ-specific variations. They can be vast: for example, watch what happens to the vascular bed of the jejunum when it's exposed to cholecystokinin:

These data are from Chou et al., and - though they are illustrative - one must acknowledge that the average ICU specialist will a) never infuse cholecystokinin into people, and b) has little interest in the specific vascular resistance of the jejunum. However, non-average ICU specialists are also clearly present in the audience, and those might be interested in all sorts of unusual things. For these weirdos, the vascular resistances of several organs are offered here in a table, derived from Karlsson et al. (2003). These were measurements taken from control animals (the treatment arm from this study on nimodipine is not shown). This also demonstrates the limitation of resistance as an index: tissues with very poor blood flow (e. the resting biceps, the abdominal skin) have preposterously high resistance values. 

Measurement and documentation of vascular resistance

Resistance, being a change in pressure divided by flow, should be measured in units of pressure per unit volume per unit time, eg. pascal seconds per cubic metre  (Pa·s/m³). However, this being medicine and physiology, anachronistic weirdness penetrates everything. Why have sensible units, when you can eponymise the name of a famous Australian cardiologist from the 1950s?

• Hybrid resistance units (HRU)s, or "peripheral resistance units" (PRU), or "vascular resistance units" (VRU), or "Wood units" after Paul Wood (1958). These describe the ratio of a driving pressure of 1 mmHg to a resultant flow of 1 L /min. These units are in mmHg⋅min⋅mL-1 and are still in use by the AHA (Kula et al, 2016).
• Dynes.s.cm-5 is the notation usually encountered at the bedside, as various measurement devices such as PA catheters typically report their data in dyne.s.cm-5. Reading critical care literature tends to make these a more familiar notation, as intensivists all over are more accustomed to looking at PA catheter data then AHA guidelines statements when it comes to vascular resistance. These can be interconverted: one Wood unit is approximately 80 dyne.s.cm-5.
• Pressure per blood flow (mmHg/ml/100g/min) is how vascular resistance to individual organs and tissues is typically reported; for example, where cerebral vascular resistance is discussed (eg. Kety et al, 1948).

And it is at this stage that every textbook rolls out the Poiseuille equation with a sort of tired certainty. It is offered as an explanation of why the resistance is as it is, whereas the old (R=P/Q) equation merely describes what it is. The Hagen-Poiseuille equation is usually stated as: 

R = (8 l η) / πr4

where 

• l = length of the vessel
• η = viscosity of the fluid
• r = radius of the vessel

As one might expect with anything borrowed from theoretical physics and shoved crudely to fit biological systems, there are some problems with applying this equation to the living circulation:

• The equation models constant flow, and that's clearly not what happens in biological systems. With pulsatile flow, the added element of kinetic energy interferes with the use of a pressure gradient in the equation.
• The equation only works for laminar flow, which cannot be uniformly expected from all the vessels. However, it appears that turbulent flow is something of a rarity in the distal arterioles. According to Ku (1997), the Reynold's number in the circulation is about 4000 in the ascending aorta at peak systole (the most turbulent time of all), falling to 600 in the abdominal aorta, 300 at the carotid bifurcation and even lower distally.
• The equation only works for cylindrical tubes, and blood vessels may not have a uniformly circular crosssection.
• The equation requires a length, and that is challenging to measure or estimate in the living circulatory system.
• The equation implies a rigid tube, whereas the arteries are elastic.
• The rheology of blood changes with vessel diameter, i.e. it becomes less viscous in smaller vessels (less than 0.5mm diameter). This is known as the Fahraeus–Lindquist effect, and is explained by the cellularity of blood, the deformability of red cells, and their aggregation.

However, for the purpose of answering exam questions, one needs to be familiar with the Poiseuille equation. There is a good chance that you will need to reproduce it in a written question or viva, and then discuss its components. It is likely that an indepth understanding will not be expected, and a passing grade will be earned by recognising that blood vessel radius is the most important determinant of peripheral vascular resistance.

Length of the blood vessels

The Poiseuille equation calls for a length value to plug in, which probably works very well in the lab with pieces of tubing. The complexity of the human circulatory system makes this value difficult to discuss, to say nothing of obtaining measurements or estimates. Of course, often one comes across some vulgar clickbait gibberish (eg. "Surprising Facts About the Circulatory System") where totally insane statements are made with no scientific basis ("...That means a person's blood vessels could wrap around the planet approximately 2.5 times!", and so on). Needless to say, this is not the sort of thing you would want to spout during a CICM physiology viva.

Outside of popular science journalism, the physiology world has a long rich history of trying to figure this out. All that "100,000 km of blood vessels" stuff probably comes from some estimates made by Harold Green (1944) on the basis of even older histological measurements made by Franklin Mall (1887). They were working from the approximate dimensions measured in fixed sections of dog mesentery; knowing the number and length of branches allowed them to make educated guesses with regards to the number (and therefore total length) of vessels of any given diameter. Complaining that formaldehyde shrinkage might have introduced errors into the measurements taken from fixed samples, Mary Wiedeman (1962) generated more accurate estimates by scrutinising the transparent wing skins of live unanaesthetised bats. Unfortunately, no mention is made of how many bats were required or what steps the author took in handling them, but what we do know is that she went on to be known among her colleagues as "the bat woman",  and ultimately received the Malpigh film award from the European Microcirculatory Society in 1979, for "a film that showed the action of blood flow in a bat's wing after thrombosis had been induced."  That film is not available to the casual Googler. However, it would be useful to use an image to impress a sense of scale upon the reader. The slighly modified picture below comes from Segal et al (2005); it is the microcirculatory network of a mouse gluteus. 

So, this is the most important section of the circulation, as that seems to be where all the resistance happens. In case anybody is interested in bat measurements, the average length of these vessels was about 3.5mm (we are talking about vessels at the bottom of the arteriolar range, with a diameter of around 15 μm). This figure is clearly going to be different for non-bats, and other authors reported values in a wide range. Unfortunately, it was impossible to track down the exact origin of these figures, but they come from an authoritative source (McDonald's Blood Flow In Arteries). This table (Table 2.2, p.36 of the 6th edition) gives some estimates of vessel dimensions from a large 20kg dog:

Thus, theoretically, using these numbers one could pick a typical arteriole, estimate the flow through it, then calculate the resistance of that solitary arteriole, and then apply the usual equation to solve for RT (the total resistance) in a  parallel resistor array,

1/RT = 1/R1 + 1/R2 + 1/R3 + ... 1/Rn,

Where n = about 2,800,000 or however many arterioles there are. 

Thus:

RT = 1 / (1/R1 + 1/R2 + 1/R3 + ... 1/Rn)

This would obviously be quite a pointless exercise, as it could never be expected to yield an accurate resistance value.  Nor does it really matter what the length of these vessels is, if one thinks about it. Autoregulatory control of peripheral vascular resistance is not carried out by varying the length of the vessels.  It will suffice to say that in most authoritative sources, length is mentioned and then politely ignored, as it is not a variable which changes in the normal course of circulatory function. For the purposes of the CICM First Part Exam, it is enough to know that the total resistance of parallel resistors is always going to be lower than the resistance of any individual resistor in the system, and adding more resistors into the system will result in a decrease in the total resistance.

Viscosity of the blood

Viscosity, for fluids, is a measure of the fluids' resistance to deformation at a given flow rate. It represents the internal friction within the fluid itself, where parts of the fluid are in motion relative to other parts. Conventionally, this property is measured using a viscometer, where the flow of blood through a narrow tube is measured. Unfortunately, being forced to travel through an unnatural glass tube is in itself something that increases the viscosity of the blood. When Whittaker & Winton (1933) compared in vitro viscosity of blood with its viscosity while flowing along the friendly endothelial surface of a dog hindlimb, the difference was startling, particularly as the haematocrit increased. The graph from the 1933 publication is reproduced here with absolutely no modification, as it is clear enough in its original form:

This brings us to the next point.  Even acellular plasma is pretty viscous, and adding more cells to it gives rise to an increase in the viscosity, as more cells means more internal friction. Moreover, whole blood has some weird properties:

• Under conditions of low shear stress, the apparent viscosity increases markedly. 
• Under high shear stress (i.e in small vessels,) the viscosity is decreased.

Plasma is generally described as either a proper godfearing Newtonian fluid, or slightly non-Newtonian. What that means is that in viscometer experiments, over a range of shear stress levels, the viscosity remains the same (or changes only slightly). Plasma obviously still has particles in it, but because they are quite small in comparison to the size of the capillary, their presence in the lumen does not disturb the orderly Newtonian behaviour. For example, the largest protein in human plasma is usually the fibrinogen molecule, and that is only about 50 nm in length, which is about 1% of the capillary diameter (Weyland, 1967).

However, red cells are about the same size as the capillary, and their presence in the bloodstream changes the viscosity significantly. Litwin & Chapman (1970) report that viscosity increases by 50% between a haematocrit of 36% and a haematocrit of 53%. More interestingly, the presence of red cells has implications for blood viscosity in capillaries. In short, blood has greatly decreased viscosity in small vessels, a phenomenon known as the Fahræus–Lindquist effect. Here's the findings from the original 1931 paper by Robin Farhaeus and Torsten Lindquist. 

This effect is attributed to the central migration of cells in the flowing blood column; they cit in the centre and never touch the sides. The outside of the column, which comes into contact with the endothelium and glycocalyx, is made up entirely of plasma, and therefore viscosity is decreased there. This idea is called Haynes’ marginal zone theory, after Robert Haynes who proposed it in 1960. To be fair, there are actually a whole range of similar marginal theories, all of which agree on the basic principles but not on how to calculate the finer details, eg. how the thickness of the marginal cell-free layer changes with changing haematocrit, etc. 

Anyway. In summary, for the purposes of the exam, it would be useful to know a handful of factors which affect viscosity. These would be:

• Vessel diameter: viscosity decreases in small vessels
• Haematocrit: viscosity increases with raised haematocrit
• Protein content of plasma: the higher the protein content (esp. fibrinogen), the greater the viscosity
• Lipid content of blood
• Plasma expanders, eg. gelatin or starch 
• Temperature - viscosity increases at low temperatures and is markedly increased below 27°C (Litwin & Chapman, 1970).

As you can see, the vessel calibre at which viscosity starts to drop is also the calibre of the arterioles most responsible for most of the resistance the arterial circulation, i.e. below 300-200 μm. From this, we can infer that blood viscosity plays a fairly minor role in peripheral vascular resistance. Moreover, it is not expected to change significantly over short timeframes, and it is a poor candidate for regional or systemic autoregulation (i.e. no reasonable organism would try to regulate their microcirculatory flow by making rapid regional adjustments to the viscosity of their blood). This leaves radius as the only meaningful Poiseuilleian variable.

Radius of the vessel

Arterioles with a radius of 100 μm and under are muscular structures and respond to a range of stimuli by constricting and relaxing, a phenomenon which can take place over a timeframe of seconds.  This change in crossectional area is the main mechanism by which systemic and regional vascular resistance is regulated. It is especially powerful because the fourth power of the radius is factored into the Poiseuille equation. In other words, a 50% decrease in radius yields a 94% decrease in blood flow, all other parameters remaining equal.

Arteriolar diameter is in fact a constantly changing dynamic variable. It undergoes a normal cyclical relaxation and contraction, which is an underappreciated phenomenon (insofar that the author had not appreciated it prior to writing this). It is slow in larger vessels (2-3 cycles every minute), becoming more frequent in the distal circulation (10-20 cycles per minute).  Meyer & Intaglietta (1986) found that this normal rhythmic movement results in a 10% diameter variation among larger 100-150 μm arterioles, and reaches 100% in the terminal arterioles, i.e. they close completely during their normal cycle. That gives one some sort of an impression of the dynamic range one can expect. 

They also change their dimensions quite rapidly. For example, Honig et al. (1980) found that muscle arterioles could dilate to recruit blood flow from nil to maximum over the span of fifteen seconds (flow increased seven-fold, representing a doubling of the vessel radius). Constriction is even faster. Here, from one of the first papers documenting the use of a video microscope to measure the microcirculation, is the effect of a few drops of noradrenaline being splashed over a frog mesentery:

So, this is clearly where the money is, in terms of exam revision. Factors which determine arteriolar radius are the factors which determine peripheral vascular resistance.  For the purpose of the ensuing discussion (and to satisfy the natural urge for classification), these can be separated into systemic and regional factors.

Systemic determinants of peripheral vascular resistance

Arterial baroreflex control is covered beautifully by Kiichi Sagawa (2011), but most people will never get to read it because it has been paywalled by Wiley. In summary, these are the main mechanism of homeostatic regulation of blood pressure, and they achieve this control at least in part because of the modulation of peripheral arteriolar smooth muscle tone:

• These are pressure receptors located in the carotid sinus and aortic arch
  • Aortic arch receptors are innervated by the aortic nerve, a branch of the vagus
  • Carotid sinus receptors are innervated by the sinus nerve of Hering, which is a branch of the glossopharyngeal nerve
  • Both synapse within the nucleus tractus solitarius
• Increased arterial wall stretch increases the firing frequency of these receptors
• Activation of these receptors leads to a decrease in sympathetic tone, which decreases both peripheral vascular resistance and the cardiac output. In other words, an increase in the receptor firing rate ultimately has an inhibitory effect on vasomotor tone.

Cardiovascular chemoreceptor reflexes,  though normally involved mainly in the central control of ventilation, can influence the peripheral vascular resistance by means of the autonomic nervous system. The normal cardiovascular response to hypoxia is to increase cardiac output and systemic peripheral resistance, something that contributes to the chronic hypertension in obstructive sleep apnoea (Trzebski, 1992).

Pulmonary baroreflex control is very similar to the arterial baroreceptor control, except the receptors are in the pulmonary arteries, and stimulating them has the opposite effect. When they are stimulated by strecth, these receptors lead to an increase in sympathetic activity - a higher peripheral vascular smooth muscle tone and a higher cardiac output (Simpson et al, 2020). In summary:

• Pulmonary arterial walls contain baroreceptors 
• Stimulation of these receptors results in reflex systemic vasoconstriction.
• They also adjust the homeostatic setpoint of the arterial baroreceptors
• The net effect is to increase the sympathetic response to hypoxia (as hypoxia increases pulmonary artery pressure and stimulates the reflex)

Hormonal influences are numerous. Apart from ones you can immediately bring to mind (angiotensin, vasopressin, etc), there are numerous endocrine mediators which in one way or another influence the systemic vascular resistance.

Temperature clearly has an effect on peripheral vessel tone. Heat causes vasodilation, and cold causes vasoconstriction. This is generally viewed as a sympathetic reflex and forms an essential part of systemic thermoregulation (Alba et al, 2019). It is hard to know whether to classify it as a systemic or regional phenomenon; a large region of skin needs to be cool before a whole-body cutaneous vasocontrictor response is elicited. On top of this systemic effect, local effects also exist (see below)

Regional factors which affect peripheral vascular resistance

This information is mainly summarised from Clifford (2011), whose article contains a massive bibliography of supporting references. The discussion here is limited to factors which are regional in origin, and which have local effects on the blood flow of one specific organ or tissue area. This was the topic of the very vaguely worded Question 12 from the first paper of 2014, and is discussed in greater detail in the chapter which summarises regonal circulatory systems. 

Myogenic regulation is regulation in response to muscle stretch, orchestrated by the smooth muscle itself (hence "myogenic"). The process can be summarised as follows:

• First, the increased pressure inside the vessel stretches the vascular smooth muscle
• The smooth muscle depolarises in response to the stretch (by action of some sort of mechanosensitive ion channel, we think - but nobody actually knows)
• This opens voltage-gated calcium channels and produces and intracellular calcium increase
• The increase in intracellular calcium is the final common pathway of all vasoconstrictor stimuli, as it causes myosin light chain phosphorylation.

This mechanism seems to have different potency depending on which organ and tissue you are looking at; for example cerebral arterioles are much more briskly responsive than the arterioles of skeletal muscle. 

Metabolic regulation can be loosely described as "regulation of blood flow which is determined by metabolic demand"; i.e. the arterioles dilate in order to promote flow to the organ or tissue which suddenly requires more metabolic substrate. This mechanism is occasionally referred to as "active hyperemia" or "functional hyperemia". Realistically, the dilation seems to occur not to meet the increased demand for metabolic substrate, but as a response to the release of metabolic endproducts by the tissue, particularly the products of anaerobic metabolism released due to ischaemia. There are a large number of substances answering that description. A good list is offered by R.J Korthius (2011):

• Carbon dioxide
• Lactate
• Potassium ions
• Adenosine
• Nitric oxide
• Carbon monoxide
• Hydrogen sulfide
• Hydrogen peroxide

Flow or shear-associated regulation is the phenomenon of proximal vasodilation in response to distal vasodilation. To put it in a less confusing form:

• Local vasodilation in small distal arterioles results in increased flow to that specific little chunk of tissue.
• As a result, the larger proximal arteriole feeding that chunk of tissue experiences an increase in flow.
• This increase in flow produces increased shear stress on the walls of the larger proximal arteriole.
• This shear stress promotes the release of various vasodilatory mediators from the affected endothelium and produces vasodilation of the larger proximal arteriole.

This response is relatively sluggish, as far as vasomotor reflexes go. According to Clifford (2011), it takes 30-40 seconds to reach a new baseline after a maximum distal stimulus has been applied.

Conducted vasomotor responses represent regional control of one region by the vasomotor events of another neighbouring region. This is mediated by the electrotonic propagation of signals from one smooth muscle cell to another. Both vasodilation or vasoconstriction can propagate in this manner.

Immunological regulation: the release of vasodilating substances in the setting of infection or inflammation is clearly an important mechanism of regional vasodilation, and is discussed in detail elsewhere. In summary, these mediators include histamine, serotonin, prostaglandins E1 and E2, etc. etc.

Local cooling can produce both regional vasoconstriction and vasodilation (Alba et al., 2019). Vasoconstriction seems to be mediated by the rho kinase pathway and is unrelated to the autonomic nervous system. Cold-induced vasodilation is also apparently a phenomenon intrinsic to the local vascular smooth muscle; it is characterised by a late increase in blood flow to regions of cool skin, and appears to be protective against frostbite (in the sense that populations whose cold-induced vasodilation responses are limited, such as African-Americans, are more prone to frostbite).