Regulating Arterial Pressure

Cut the nerves that report arterial pressure to the brainstem of a conscious dog, and, after the first days, the mean pressure is little changed. Cut them, and the pressure swings wildly from minute to minute. One reflex, two results: the arc is the fastest controller of arterial pressure, and on this evidence not the main thing that sets its level, though a chronic contribution is still argued. Mean arterial pressure (MAP) in an adult human at rest is about 93 mmHg for a pressure of 120/80 mmHg and, in Guyton’s summary, rarely strays more than 10 to 15 percent through a day. Three controllers hold it there, on different clocks.

Stretch in the carotid sinus becomes a slower heart and wider arterioles

Baroreceptors are stretch receptors in the walls of the two carotid sinuses, where each common carotid artery divides, and in the aortic arch; they sense not pressure but the deformation it causes in the wall around them. Low-pressure cardiopulmonary receptors in the atria, ventricles and great veins report central blood volume instead. In mice, PIEZO1 and PIEZO2 are together required for baroreception (vagal sensory neurons, mechanotransduction channels).

Carotid sinus afferents run in a branch of the glossopharyngeal nerve, cranial nerve IX, which Hering identified in the 1920s; aortic arch afferents run in the vagus. Both end in the nucleus of the solitary tract, which excites the caudal ventrolateral medulla, which in turn inhibits the rostral ventrolateral medulla, the source of tonic sympathetic drive. The same input excites vagal motor neurons in the dorsal motor nucleus and nucleus ambiguus. Autonomic anatomy and transmitters are in the hypothalamus and autonomic control; the carotid body chemoreceptors at the same bifurcation are in control of breathing.

A rise in arterial pressure therefore raises afferent firing, raises vagal output and lowers sympathetic output: the interbeat interval lengthens and heart rate falls, arteriolar smooth muscle relaxes and total peripheral resistance falls (smooth muscle), and venous tone falls, lowering mean systemic filling pressure and venous return (cardiac output and venous return). A fall in pressure reverses every arrow. The reflex works through all three effectors because MAP has more than one determinant, and their relative contributions change with timing and conditions:

MAP = cardiac output x total peripheral resistance, and from a pressure trace, MAP is about diastolic pressure + pulse pressure / 3

Both come from resistance, compliance and local flow and are used here, not derived again.

Seconds, minutes, days: three controllers and their gains

Guyton’s 1991 summary sorts the controllers by their clocks: neural receptors act within seconds, hormonal controllers within minutes, a kidney-fluid system over hours to days. A fast controller is judged by how large an error it leaves, not by whether pressure is normal. Gain is the measure: at steady state, the correction divided by the error that remains, extending the threshold-and-gain vocabulary of the hypothalamus and autonomic control. Rearranged, the remaining error equals the open-loop disturbance divided by one plus the gain.

Worked problem, part A: how much of a disturbance does the reflex remove?

Illustrative round numbers, not measured values. Assumptions: an adult human at rest, steady state, linear relations, and a step rise in total peripheral resistance as the disturbance.

1. Baseline. MAP is 95 mmHg. With the reflex cut, the disturbance would take it to 115 mmHg: an open-loop change of +20 mmHg.
2. Close the loop. With the reflex intact MAP settles at 101.7 mmHg, so the remaining error is 6.7 mmHg and the correction is 20 - 6.7 = 13.3 mmHg.
3. Gain. 13.3 / 6.7 = 2.0.
4. Check. Remaining error = open-loop change / (1 + G) = 20 / 3 = 6.67 mmHg, which matches.

A gain of 2 removes two thirds of a fast disturbance; with the reflex gone the whole 20 mmHg stands. And baroreceptors reset, shifting their firing range toward the prevailing pressure, so a steady error is reported more weakly. Resetting can be incomplete, so a sustained change in pressure can still alter reflex output.

Cut the nerves and the mean rises, then settles back

Does the arc set the daily mean? Cowley, Liard and Guyton tested it in 1973 in dogs, comparing pressure before and several days after sinoaortic denervation, section of the carotid sinus and aortic depressor nerves. If the reflex sets the level, the mean should settle somewhere new and stay. As later reviews summarize the paper, the mean pressure was only minimally increased, while fast variability rose dramatically, with large swings on posture change and activity. The authors concluded that the reflex minimizes variations rather than setting the chronic level.

Ruled in: the baroreflex is the main buffer of second-to-minute swings. Not supported, for this dog preparation: the arc as setter of the daily mean. Not shown: that it has no chronic role, since cutting a nerve invites central reorganization, and the rise that does follow denervation in dogs is described as largely transient and attributed to that plasticity. In Thrasher’s 2005 telemetry study, dogs after sinoaortic denervation averaged 124 mmHg in the first week against a control of 94, 103 in the second, and values not different from control in weeks 3 to 5. Species differ: Krieger’s group reported chronic increases in rats, and in telemetered normotensive Wistar-Kyoto rats the mean rose above sham for several days, with none in spontaneously hypertensive rats, while the standard deviation rose in both strains. The clean statement is about lability, not level. Cowley’s 1992 review generalized it: baroreceptors reset and cannot hold a sustained signal, so neural control is limited to correcting rapid changes.

Then the experiment was run the other way around. Lohmeier, Irwin, Rossing, Serdar and Kieval drove electrodes around both carotid sinuses of conscious instrumented dogs from a pulse generator for seven days. Control mean pressure was 93 plus or minus 3 mmHg at 64 plus or minus 4 beats per minute; stimulation took it to 75 mmHg on day 1 and the fall held all seven days. Plasma renin activity did not rise despite the lower pressure, which the authors read as sympathetic inhibition of renin secretion.

Ruled in: afferent input can lower pressure for days here. Not shown: that the natural reflex does so, since electrodes bypass resetting at the nerve ending; that it works without the kidney, since, if salt intake was constant and balance was restored, a sustained fall means the renal relation below moved; or anything beyond normotensive dogs over a week. The mirror experiment agrees: Thrasher ligated the common carotid artery below the one intact carotid sinus in dogs whose aortic and other carotid sinus nerves had been cut, and pressure rose about 22 mmHg over the week and normalized when flow was restored; in a five-week repeat the rise waned to about 10 mmHg above control.

Neither side is settled. The baroreflex is the dominant short-term controller; whether it contributes to the chronic level is open, with dog experiments pointing both ways.

The kidney sets the level: pressure natriuresis

Pressure natriuresis is the rise in renal sodium excretion when renal perfusion pressure rises (glomerular filtration, tubular transport). Plot excretion against mean arterial pressure and you have the renal function curve; sodium intake is a horizontal line across it. They cross at one pressure, the long-term set point: anywhere else, intake and output differ, extracellular volume changes, and pressure moves. That is Guyton’s infinite gain argument. The fast reflex is proportional and leaves an error; the kidney is an integrator, changing volume while any error remains, so for a given curve and intake the steady-state error is zero. The testable corollary: every other controller can set a chronic pressure only by moving the curve.

Worked problem, parts B and C: where the kidney puts the pressure

Illustrative linear curves, not fitted to data. Assumptions: steady state, intake equals excretion at equilibrium, curve unchanged unless stated.

Part B. Renal function curve: excretion = 10 mEq/day per mmHg x (MAP - 85 mmHg). Intake 150 mEq/day.

1. Equilibrium. 150 = 10 x (MAP - 85), so MAP = 100 mmHg.
2. Apply a resistance step. Part B is a separate illustration with its own baseline and a smaller step than part A: MAP rises acutely by 10 mmHg, to 110 mmHg, on an unchanged curve, where excretion is 10 x (110 - 85) = 250 mEq/day, a net loss of 100 mEq/day.
3. Consequence. Extracellular volume falls, mean systemic filling pressure and cardiac output with it, and MAP returns to 100 mmHg, the only pressure where output equals intake. Remaining error zero, over days rather than instantly.

Part C. Move the curve instead. (i) Shift it 10 mmHg right, excretion = 10 x (MAP - 95): equilibrium MAP = 95 + 15 = 110 mmHg, sustained, at the same intake. (ii) Flatten it through the same normal point, excretion = 5 x (MAP - 70), which also gives 100 mmHg at an intake of 150. At 250 mEq/day the normal curve gives 110 mmHg, a rise of 10; the flat curve 120 mmHg, a rise of 20. A shallower curve means a larger pressure change per increment of salt, which is what salt sensitivity means. (iii) Holding mean pressure 18 mmHg below control for a week at unchanged intake needs the curve about 18 mmHg left, at unchanged slope, once intake and output are balanced again, so a sustained fall of this kind would require some change in how the kidney handles sodium; whether the 2004 dogs reached balance is not stated here.

Disturbance D, open-loop rise in MAP 20 mmHg Baroreflex gain G 2 Renal curve shift 0 mmHg (+ right, higher pressure; − left) Sodium intake 150 mEq/day

Lesson parts A and B values Lesson part C(i): shift +10

Remaining error = D / (1 + G)

Correction = D − remaining error

Gain = correction / remaining error

Kidney: excretion = 10 mEq/day per mmHg × (MAP − intercept) = intake

In this sketch: an illustrative model, not fitted to data, for an adult human at rest in steady state with linear relations and a step rise in total peripheral resistance as the disturbance. The lesson’s own illustrative numbers are the fast baseline of 95 mmHg (part A) and the renal slope of 10 mEq/day per mmHg with an intercept of 85 mmHg (part B). Made up for the sketch: the slider ranges and extending the straight lines beyond the worked values. The lesson gives the fast bars and the kidney bar as separate illustrations (baseline 95 mmHg against a kidney equilibrium of 100 mmHg at an intake of 150), so compare how they move, not their offsets. Baroreceptor resetting is left out.

A reflex of gain G shrinks a disturbance to D / (1 + G) but never to zero, while the pressure the kidney settles at is where excretion equals intake, so it moves only when the curve or the intake moves.

The hormones are how the curve gets moved. Renin is an enzyme from juxtaglomerular cells around the afferent arteriole; three inputs raise its release, a fall in perfusion pressure, a fall in NaCl delivery sensed at the macula densa, and beta-1 adrenergic input, while angiotensin II feeds back negatively on it. Renin cleaves liver-derived angiotensinogen to angiotensin I, and angiotensin-converting enzyme makes angiotensin II, which constricts systemic arterioles, releases vasopressin, and stimulates aldosterone secretion from the adrenal zona glomerulosa along with sodium reabsorption. Aldosterone binds the mineralocorticoid receptor and brings epithelial sodium channels (ENaC) to the apical membrane of collecting tubule cells (the adrenal glands). Vasopressin, released strongly when circulating volume falls, drives water reabsorption in the collecting duct through V2 receptors and aquaporin-2, which is its main action on volume, and constricts arteries through V1a receptors at the higher concentrations of hypovolemia, both the V1a and the angiotensin II receptor AT1 being G protein-coupled (G proteins). Pushing the other way, atrial natriuretic peptide from the atria and B-type natriuretic peptide, released mainly from the ventricles, cause diuresis, natriuresis and a fall in pressure (the kidney, water, sodium and potassium). On a high salt intake, angiotensin II and aldosterone fall, steepening pressure natriuresis, so balance is kept with little pressure change while the system works normally. In Guyton’s framework these controllers act by moving the curve rather than by a separate route to a chronic pressure.

Standing, losing volume and straining

Three perturbations, read as rises and falls. Standing up: gravity moves blood into the veins of the legs and abdomen, so venous return, stroke volume, cardiac output and arterial pressure fall for a few beats; carotid sinus stretch falls with them, so sympathetic outflow rises and vagal outflow falls, and heart rate, resistance and venous tone rise until pressure recovers. A 2026 mouse study adds cardiac vagal PIEZO2 volume receptors to this compensation; losing them gives orthostatic hypotension in that species. A fall in blood volume: venous return, cardiac output and pressure fall; arterial and cardiopulmonary receptors unload; sympathetic outflow, heart rate and total peripheral resistance rise; renin, angiotensin II, aldosterone and vasopressin rise, atrial stretch and natriuretic peptide release fall, and capillary refill from the interstitium begins (capillary exchange and lymph). A Valsalva strain, a forced expiration against a closed airway: in phase I the sudden rise in intrathoracic pressure briefly raises pressure mechanically, and the loaded baroreceptors slow the heart by reflex; in phase II, the strain lowers venous return, left ventricular dimensions, stroke volume and output, so pressure falls, then recovers late as sympathetic activity raises total peripheral resistance, with heart rate rising as pressure falls; phase III, at release, brings a brief further fall; phase IV overshoots above baseline, loading the baroreceptors, with reflex bradycardia and sympathetic withdrawal.

Why some pressures stay high

About 90 to 95 percent of human hypertension is essential, multifactorial and gene-environment in origin, with heritability estimated between 35 and 50 percent in most family and twin studies (complex traits and GWAS). Salt sensitivity is continuous rather than two bins, and in salt-sensitive people a flatter effective pressure-excretion relation describes the larger pressure response to salt; its position and its cause vary, and most salt-sensitive people do not retain more of a salt load than others. Guyton and Hall argue that in every form of human and experimental hypertension studied so far the curve sits at higher pressures, though the relation is hard to measure directly in people, while resistance increases that spare it, such as closing an arteriovenous fistula, give only transient rises. Kidney transplantation attenuates or abolishes genetic salt-sensitive hypertension in rodent strains, which points at the organ.

A monogenic form makes the logic concrete. In Liddle syndrome, gain-of-function mutations in the beta subunit of ENaC, shown by Shimkets and colleagues in 1994, or in the gamma subunit, leave more channels active in the collecting duct apical membrane, so sodium reabsorption is higher at any perfusion pressure and the curve sits to the right. Rising: sodium reabsorption and extracellular volume first, then arterial pressure, until pressure natriuresis restores balance at the new higher level; the lumen-negative potential that sodium entry creates also drives potassium secretion, so plasma potassium is often low. Falling: renin, suppressed by the high pressure and volume, and with it angiotensin II and aldosterone. That separates the mutation from primary aldosteronism, where aldosterone is high and renin low (the adrenal glands).

Design problem

The denervated dogs and the stimulated dogs appear to disagree. You have conscious dogs with arterial telemetry and electrodes around both carotid sinuses. Design the test of whether the sustained fall in pressure during stimulation depends on the renal nerves, then compare your design and prediction with what Lohmeier and colleagues found in 2007.

One way to do it

1. Groups. Intact dogs and dogs after bilateral renal denervation, each with sham-operated controls, all carrying the same electrodes.
2. Protocol. A control period, seven days of stimulation, then recovery, with continuous 24-hour pressure and heart rate, so transient and sustained effects separate.
3. Measures. Fixed sodium and water intake with daily urine collection, so sodium balance is tested rather than assumed, with further controlled intake or pressure conditions needed if the aim is to map the curve, plus plasma renin activity and noradrenaline as indirect indices of renin release and sympathetic outflow, with nerve recording or noradrenaline spillover needed for a direct sympathetic measure.
4. Predictions. If renal denervation abolishes the fall, those nerves carry the chronic effect; if it persists, something else moves the curve, such as other sympathetic beds or circulating hormones.
5. Limits. Surgically altered normotensive dogs over one week; renal nerves regrow, so confirm denervation afterwards; and the electrodes still bypass the receptor.