33. Capstone: Systems Integration

How the Body Works as One Unified Machine

Learning Objectives

33.1 The Problem With Organ Systems

You have now met all eleven of them. It is time to say plainly what has been implied since Chapter 1: the eleven-system partition is not a description of the body. It is a filing scheme, and like every filing scheme it was built for the convenience of the filer.

Its origin is dissection. When the organizing categories of anatomy were fixed — largely in the sixteenth through nineteenth centuries — the available evidence was a body on a table. Things that were physically continuous got grouped together, because physical continuity was what you could see. The stomach connects to the intestine, so they are one system. The kidney connects to the ureter, which connects to the bladder, so those are one system. This produces a map of anatomical contiguity.

Physiology is a map of functional coupling, and the two maps are not the same map.

What the partition hides

1 · Organs with two citizenships. The pancreas is a digestive organ and an endocrine gland, and the two cell populations sit within a millimetre of one another. The kidney is a urinary organ, an endocrine gland (renin, erythropoietin, calcitriol), and the principal long-term regulator of arterial pressure. Bone is a skeletal organ, the entire site of blood and immune cell production, the body's calcium and carbonate reservoir, and an endocrine gland secreting osteocalcin and FGF23. Skin performs a required step in vitamin D synthesis and is one of the largest vascular capacitance beds in the body. The heart secretes hormones. The lung is the site of a critical enzymatic step in blood pressure control. Every one of these facts is taught as a curiosity in one chapter and never used in another.

2 · Regulatory relationships that cross every boundary. Count them and the result is startling. Of the 110 ordered pairs of distinct organ systems, 101 — 92% — carry a direct regulatory influence in the direction stated. Not "they are both in the body." A specific molecule, nerve, or mechanical linkage by which one system changes the behavior of another.

3 · Variables that no system owns. This is the deepest failure. Blood pH is defended by the lung, the kidney, the blood, bone, muscle, and liver. Mean arterial pressure is defended by the heart, the vessels, the kidney, the brainstem, the adrenal cortex, the posterior pituitary, the lymphatics, and skeletal muscle. A partition has nowhere to put a shared variable, so shared variables get taught four or five separate times, in four or five separate chapters, as four or five separate topics — and the student, entirely reasonably, concludes that they are four or five separate topics.

4 · The order in which things fail. When perfusion collapses, organs fail in a strict and predictable sequence (§33.5). That sequence is a property of the network — of who is sacrificed to protect whom, and of which tissues have no extraction reserve — and it is invisible in a list.

5 · Drugs. Every drug Amara takes acts in at least two systems. Metoprolol blocks β₁ receptors on her sinoatrial node and on the juxtaglomerular cells of her kidney. Lisinopril inhibits an enzyme on her pulmonary capillary endothelium in order to change sodium handling in her distal nephron and collagen deposition in her myocardium. Empagliflozin was designed as a diabetes drug and turned out to be a heart failure drug and a kidney drug, for reasons nobody fully anticipated. A partition predicts one effect per drug. A network predicts side effects, and it predicts the good surprises too.

                    WHO DIRECTLY REGULATES WHOM
        Row = the system exerting the influence.  Column = the system
        receiving it.  ● = strong direct regulation   ○ = real but
        secondary   ·  = no direct link (influence only via a third system)

                        ── RECEIVING SYSTEM ──────────────────────────
                        IN  SK  MU  NV  EN  CV  LY  RS  DG  UR  RP
     IN  Integumentary   ─   ●   ·   ○   ●   ●   ●   ·   ·   ○   ·
     SK  Skeletal        ·   ─   ●   ○   ●   ●   ●   ●   ○   ●   ○
     MU  Muscular        ○   ●   ─   ○   ●   ●   ●   ●   ○   ○   ○
     NV  Nervous         ●   ○   ●   ─   ●   ●   ○   ●   ●   ●   ●
     EN  Endocrine       ●   ●   ●   ●   ─   ●   ●   ○   ●   ●   ●
     CV  Cardiovascular  ●   ●   ●   ●   ●   ─   ●   ●   ●   ●   ●
     LY  Lymph/Immune    ●   ●   ○   ●   ●   ●   ─   ●   ●   ●   ○
     RS  Respiratory     ·   ·   ●   ●   ●   ●   ○   ─   ·   ●   ·
     DG  Digestive       ○   ●   ●   ●   ●   ●   ●   ○   ─   ●   ○
     UR  Urinary         ○   ●   ○   ●   ●   ●   ○   ●   ○   ─   ○
     RP  Reproductive    ●   ●   ●   ●   ●   ●   ○   ○   ○   ○   ─

   EXAMPLES OF SINGLE CELLS IN THE MATRIX
     IN → SK   UVB light on skin makes cholecalciferol → calcitriol →
               intestinal Ca absorption → bone mineral
     RS → EN   Pulmonary capillary ACE converts angiotensin I to II
     MU → LY   Skeletal muscle compression is the pump that moves lymph
     UR → SK   Renal 1α-hydroxylase is the ONLY source of active vitamin D
     DG → CV   Hepatocytes make albumin (oncotic pressure) and
               angiotensinogen (the RAAS substrate)
     LY → CV   Arterial plaque is an inflammatory lesion, not a plumbing one
     SK → LY   Every lymphocyte you own was born in bone marrow

     101 of the 110 ordered pairs (92%) are filled.  The nine blanks are
     the ONLY places where "these two systems do not talk to each other"
     is an honest statement.

What replaces the partition

Not "everything affects everything." That is true, useless, and the reason integration is so often taught badly. What replaces it is a two-part model you have been assembling for twenty-seven chapters without necessarily noticing:

• A short list of defended variables — nine of them, from §1.5 — that the body will spend any amount of energy and sacrifice any tissue to hold constant. These are the goals of the system.
• A finite set of loops that defend them. Perhaps a dozen loops matter clinically. Six of them are worked completely in §33.3.

In this model, organs are nodes. Systems are conveniences. The loops are the physiology. When you can name the variable, name the loop, and name the node that failed, you can usually predict the rest of the patient — and when you cannot, it is almost always because you have missed a loop, not because you have forgotten a fact.

33.2 The Nine Variables the Whole Body Defends

In §1.5 you were given a table of nine regulated variables and told that homeostasis was the master concept. You were not, at that point, in any position to see what the table actually claims. Here it is again, and this time every defender is named.

The claim is this: no variable on this list is defended by fewer than six organ systems acting primarily, and most are touched by all eleven. That redundancy is not sloppiness. It is the reason you are alive.

        THE NINE DEFENDED VARIABLES AND WHO DEFENDS THEM
        ● = primary defender   ○ = secondary or modulating   · = negligible

                            IN SK MU NV EN CV LY RS DG UR RP   PRIMARY
                                                               DEFENDERS
  1  CORE TEMPERATURE       ●  ·  ●  ●  ●  ●  ●  ○  ○  ○  ○       7
       skin vasomotor + sweat · shivering + exercise heat ·
       hypothalamic set point · thyroid & catecholamines ·
       convective transport by blood · pyrogens move the set point

  2  BLOOD pH  (7.40)       ○  ●  ●  ●  ●  ●  ○  ●  ●  ●  ○       8
       CO2 by lung (minutes) · H+/HCO3 by kidney (days) ·
       bicarbonate + hemoglobin + protein buffers in blood ·
       bone carbonate for chronic loads · muscle lactate flux ·
       hepatic lactate clearance and the urea cycle

  3  BLOOD GLUCOSE          ·  ○  ●  ●  ●  ●  ○  ·  ●  ●  ○       6
       gut absorption + incretins · pancreatic islet ·
       hepatic glycogen + gluconeogenesis · muscle GLUT4 sink ·
       brain as obligate consumer AND hypoglycemia sensor ·
       renal SGLT2 reabsorption + renal gluconeogenesis

  4  PLASMA SODIUM          ●  ○  ○  ●  ●  ●  ○  ○  ●  ●  ○       6
       renal Na and water handling · ADH + aldosterone + ANP ·
       hypothalamic osmoreceptors + thirst · atrial volume
       receptors · dietary intake and diarrheal loss · sweat

  5  PLASMA POTASSIUM       ○  ○  ●  ●  ●  ●  ·  ●  ●  ●  ·       7
       kidney excretes 90% · aldosterone + insulin + beta-2 ·
       skeletal muscle holds 98% of body K+ · colonic secretion ·
       arterial pH shifts K+ across every membrane ·
       the heart is both a defender and the organ that dies

  6  PLASMA CALCIUM         ●  ●  ○  ○  ●  ○  ○  ○  ●  ●  ●       6
       skin makes cholecalciferol · liver 25-hydroxylates ·
       kidney 1-alpha-hydroxylates AND reabsorbs Ca ·
       parathyroid CaSR is the sensor · bone is the bank ·
       gut absorbs · pregnancy and lactation double the demand

  7  ARTERIAL OXYGEN        ·  ●  ●  ●  ●  ●  ○  ●  ●  ●  ○       8
       alveolar ventilation · V/Q matching · Hb concentration ·
       marrow production · renal erythropoietin · dietary iron,
       B12, folate · diaphragm as the pump · chemoreceptor drive

  8  MEAN ARTERIAL PRESSURE ●  ○  ●  ●  ●  ●  ●  ●  ●  ●  ●      10
       cardiac output · systemic vascular resistance ·
       renal pressure natriuresis (the ONLY infinite-gain term) ·
       baroreflex · RAAS + ADH + ANP · splanchnic capacitance ·
       skeletal muscle pump · lymphatic return of filtered fluid ·
       skin thermoregulatory shunt · hepatic albumin

  9  PLASMA OSMOLALITY      ●  ○  ○  ●  ●  ●  ·  ○  ●  ●  ○       6
       ADH from posterior pituitary · renal collecting duct
       aquaporins · hypothalamic osmoreceptors within 1% ·
       thirst · drinking and gut absorption · hypotonic sweat

  MINIMUM number of PRIMARY defenders for any variable  . . . . . . . .  6
  Number of variables defended by a SINGLE system . . . . . . . . . . .  0

Why so many defenders? Because they act on different timescales

Redundancy in engineering usually means duplicate parts. Physiological redundancy is different and much cleverer: the defenders of a variable are layered in time, and each layer buys time for the next.

Read the column headings. The fast layers are weak; the strong layers are slow. That trade-off is unavoidable, because a mechanism that acts in one second cannot also move kilograms of salt, and a mechanism that moves kilograms of salt cannot do it in one second. The body's solution is to stack them, so that each layer holds the variable inside its tolerance long enough for the next layer to engage.

This also tells you how to read an abnormal number. When you find a variable out of range, you are seeing the residue after every defender has already done what it can. Amara's potassium of 4.6 mEq/L is not a passive fact. It is the equilibrium reached between a loop diuretic driving potassium out, an ACE inhibitor holding it in, a kidney at 46 mL/min that excretes it less well than it used to, and skeletal muscle acting as a 3,000-mEq buffer. Four opposing forces, one normal number. Change any one of them and the number moves.

The four moves

Every homeostatic response in this book is one of exactly four moves, and naming the move is often more useful than naming the hormone.

1. SHIFT — move the substance between compartments without changing total body content. Insulin driving potassium into cells. Calcium leaving bone. Blood leaving the splanchnic veins to fill the central circulation. Shifts are fast and reversible and they change the measured value without solving anything.
2. BUFFER — bind or release the substance chemically so the free concentration barely moves. Bicarbonate and hemoglobin for H⁺; albumin for calcium and for drugs; bone carbonate for a chronic acid load.
3. PRODUCE or EXCRETE — change total body content. Only a few organs can do this: the kidney (water, Na, K, H⁺, Ca, phosphate), the lung (CO₂), the gut (absorption and fecal loss), the liver (glucose, urea, albumin, clotting factors), the marrow (red cells), the skin (water and Na in sweat, vitamin D).
4. RE-SET DEMAND — change how much of the variable is needed. Fever raises the temperature set point. Shivering is suppressed in hypovolemia. Anorexia in acute illness reduces the digestive and metabolic load. Reducing heart rate with a β-blocker reduces myocardial oxygen demand rather than increasing supply — a therapeutic version of the same move, and the reason it works when nothing else does.

33.3 The Great Integrative Loops

Six loops. If you own these six completely — not recognize them, own them, forwards and backwards — you can reason about most of what happens to most patients. Everything else is detail hung on these frames.

Each is presented the same way: the variable being defended, the sensor, the chain of messengers with the organ that supplies each, the effectors, and the cost.

(a) The renin–angiotensin–aldosterone system

The RAAS is the single best example of integration in the human body, because it is physically impossible to describe without using five organs. It is also the loop most heavily targeted by drugs, and Amara takes four medications that touch it.

The variable defended: effective circulating volume — which is not the same as blood volume, and the difference is the reason heart failure is confusing. Effective circulating volume is the adequacy of arterial filling as sensed by the kidney. A patient can be fifteen litres fluid-overloaded and still have a kidney that believes the body is hemorrhaging, because what the kidney senses is perfusion, not volume.

The sensors — three of them, in one organ, any of which can fire alone:

1. The afferent arteriolar baroreceptor. Granular (juxtaglomerular) cells in the wall of the afferent arteriole are stretch sensors. Less stretch — meaning lower renal perfusion pressure — means more renin.
2. The macula densa. A patch of specialized cells in the thick ascending limb, sitting against its own glomerulus, sensing luminal NaCl delivery. Less NaCl arriving means more renin. (This is the sensor that makes loop diuretics and SGLT2 inhibitors behave in opposite directions.)
3. Renal sympathetic nerves acting on β₁ receptors on the same granular cells. More sympathetic outflow means more renin, independent of pressure or sodium.

The chain — and note the organ that supplies each step:

• Kidney releases renin, an enzyme, not a hormone in the classical sense.
• Liver continuously secretes angiotensinogen, a 452-amino-acid α₂-globulin, into plasma. It is always there in excess; renin is the rate-limiting step.
• Renin cleaves angiotensinogen to angiotensin I — biologically inert.
• Lung — specifically the enormous surface of pulmonary capillary endothelium — carries angiotensin-converting enzyme, which removes two more residues to make angiotensin II. The lung is in the blood pressure business because it is where all the blood goes, every minute, past the largest endothelial surface in the body.
• Angiotensin II acts on AT₁ receptors in at least six places at once.
• Adrenal cortex, zona glomerulosa, releases aldosterone.
• Kidney again: aldosterone enters principal cells of the collecting duct, and over 1–3 hours increases the number of epithelial sodium channels (ENaC) in the apical membrane and Na⁺/K⁺-ATPase pumps in the basolateral membrane.

The effects of angiotensin II, all six:

That last row is the one students skip and clinicians care most about. Angiotensin II and aldosterone are not only pressure hormones; they are growth factors for connective tissue. Sustained exposure lays down interstitial collagen in the myocardium, which stiffens the ventricle in diastole — and a stiff ventricle is Amara's actual disease.

The brake: the heart is itself an endocrine organ. Atrial stretch releases atrial natriuretic peptide; ventricular wall stress releases B-type natriuretic peptide. Both oppose the RAAS at nearly every point: they dilate arterioles, promote natriuresis, inhibit renin and aldosterone release, and antagonize fibrosis. Amara's NT-proBNP of 410 pg/mL is her ventricle shouting the counter-message — and being outvoted.

              THE RAAS: FIVE ORGANS, ONE LOOP, SIX DRUG TARGETS

   ┌──────────────────────────────────────────────────────────────────┐
   │ (1) KIDNEY — juxtaglomerular granular cells                      │
   │     THREE INDEPENDENT TRIGGERS:                                  │
   │       · low afferent arteriolar stretch (low perfusion pressure) │
   │       · low NaCl at the macula densa                             │
   │       · beta-1 sympathetic drive        ◄══ [A] BETA-BLOCKER     │
   │                          │                   metoprolol          │
   │                          ▼  RENIN                                │
   └──────────────────────────┼───────────────────────────────────────┘
                              │
   ┌──────────────────────────▼───────────────────────────────────────┐
   │ (2) LIVER — secretes ANGIOTENSINOGEN continuously (substrate)    │
   │            renin cleaves it  ───────────►  ANGIOTENSIN I (inert) │
   └──────────────────────────┬───────────────────────────────────────┘
                              │
   ┌──────────────────────────▼───────────────────────────────────────┐
   │ (3) LUNG — pulmonary capillary endothelial ACE                   │
   │            ANG I ──────────────►  ANGIOTENSIN II                 │
   │                     ◄══ [B] ACE INHIBITOR  lisinopril            │
   │            (ACE also degrades bradykinin → hence the cough)      │
   └──────────────────────────┬───────────────────────────────────────┘
                              │  AT-1 receptor  ◄══ [C] ARB (losartan)
        ┌────────┬────────────┼────────────┬──────────────┐
        ▼        ▼            ▼            ▼              ▼
   ARTERIOLE  EFFERENT   (4) ADRENAL   PITUITARY     HEART / KIDNEY
   constrict  ARTERIOLE     CORTEX      ADH out      FIBROBLASTS
   ↑ SVR      constrict     aldosterone  ↑ water     collagen laid down
   ↑ afterload GFR held        │         retained    → STIFF VENTRICLE
                               │  ◄══ [D] MRA spironolactone
                               ▼
   ┌──────────────────────────────────────────────────────────────────┐
   │ (5) KIDNEY again — collecting duct principal cell                │
   │     aldosterone → more ENaC apically, more Na/K-ATPase basally   │
   │     Na+ REABSORBED · K+ and H+ SECRETED                          │
   │     ◄══ [E] LOOP DIURETIC furosemide blocks NKCC2 UPSTREAM       │
   │            (lowers volume — but RAISES renin. It fights the loop │
   │             by feeding it.)                                      │
   │     ◄══ [F] SGLT2 INHIBITOR empagliflozin blocks proximal Na-    │
   │            glucose uptake → MORE NaCl reaches the macula densa   │
   │            → adenosine → afferent constriction → lower glomerular│
   │            pressure. It quiets the loop by restoring its sensor. │
   └──────────────────────────┬───────────────────────────────────────┘
                              ▼
                    PLASMA VOLUME ↑ → PRELOAD ↑ → the heart
                              │
                              ▼
        COUNTER-LOOP:  atrial stretch → ANP;  ventricular stress → BNP
                       → natriuresis, vasodilation, renin inhibition,
                         anti-fibrosis.  Amara's NT-proBNP 410 pg/mL.
                              │
                              └──────► back to (1): the kidney senses
                                       perfusion, not volume, and can
                                       call for MORE volume in a body
                                       that is already overfilled.

(b) The baroreceptor reflex

Where the RAAS operates over hours to days, the baroreflex operates in less than one second, and it is the reason you do not faint every time you stand up.

Sensors: stretch-sensitive nerve endings in the carotid sinus (at the bifurcation of each common carotid) and in the aortic arch. They are rate and level detectors: they fire in proportion to how stretched the wall is and how fast the stretch is changing.

Afferents: carotid sinus fibres travel in the glossopharyngeal nerve (CN IX); aortic arch fibres travel in the vagus (CN X). Both terminate in the nucleus of the solitary tract in the medulla.

Integration: the NTS excites the caudal ventrolateral medulla, which inhibits the rostral ventrolateral medulla, which is the tonic source of sympathetic outflow. The NTS also excites the nucleus ambiguus and dorsal motor nucleus, the source of cardiac vagal outflow. So high pressure produces sympathetic withdrawal and vagal activation simultaneously — the two arms move in opposite directions, which doubles the reflex gain.

Effectors: sinoatrial node (β₁ and M₂ — rate), ventricular myocardium (β₁ — contractility), arterioles (α₁ — resistance), veins (α₁ — capacitance, and therefore preload), and the adrenal medulla and juxtaglomerular cells for the slower reinforcement.

Cost and limitation. The baroreflex resets. Hold a person's pressure at 168/98 for three years, as Amara's was, and the reflex begins defending 168/98 as though it were normal. This is why the baroreflex cannot be the long-term regulator of blood pressure — a controller that adopts whatever value it is given has no long-term set point at all. Long-term arterial pressure is set by the kidney, through pressure natriuresis: the relationship between arterial pressure and sodium excretion, which alone among the mechanisms has effectively infinite gain, because as long as pressure is above the level at which sodium balance is achieved, the kidney keeps removing volume, indefinitely.

     THE BARORECEPTOR REFLEX — four systems, under one second

   VESSEL          NERVE            BRAINSTEM              HEART/VESSEL
   ══════          ═════            ═════════              ════════════

   Carotid sinus ──► CN IX ──┐
   (stretch ↑)      glosso-  │
                    pharyng. │      ┌──────────────┐
                             ├─────►│  NUCLEUS of  │
   Aortic arch ───► CN X ────┘      │ the SOLITARY │
   (stretch ↑)      vagus           │  TRACT (NTS) │
                                    └──────┬───────┘
                                           │
                    ┌──────────────────────┴──────────────────────┐
                    ▼                                             ▼
            ┌───────────────┐                            ┌────────────────┐
            │ CVLM excited  │                            │ NUCLEUS        │
            │      ↓        │                            │ AMBIGUUS       │
            │ RVLM INHIBITED│                            │ excited        │
            └───────┬───────┘                            └────────┬───────┘
                    │ sympathetic outflow ↓                       │ vagal ↑
                    ▼                                             ▼
      ┌─────────────────────────────┐            ┌────────────────────────┐
      │ ARTERIOLES  α1 ↓ → dilate   │            │ SA NODE  M2 → rate ↓   │
      │ VEINS       α1 ↓ → capacity↑│            │ (acetylcholine opens   │
      │ SA NODE     β1 ↓ → rate ↓   │            │  K+ channels; onset in │
      │ MYOCARDIUM  β1 ↓ → force ↓  │            │  ONE heartbeat)        │
      │ KIDNEY      β1 ↓ → renin ↓  │            └────────────────────────┘
      └─────────────────────────────┘
                    │                                             │
                    └──────────────► MAP FALLS ◄──────────────────┘
                                        │
                                        └──► baroreceptors re-sense
                                             = NEGATIVE FEEDBACK, latency <1 s

   WHAT IT CANNOT DO:  the reflex RESETS over days to whatever pressure it
   is repeatedly exposed to.  A controller that adopts its own input as its
   set point has NO long-term set point.  Long-term MAP is set by RENAL
   PRESSURE NATRIURESIS, the only mechanism in the body with infinite gain.

(c) Oxygen delivery — the most useful equation in the book

Everything the cardiorespiratory system does can be written in one line, and then that line can be taken apart until every term belongs to an organ.

DO₂ = CO × CaO₂ — oxygen delivery equals cardiac output times arterial oxygen content.

VO₂ = DO₂ × E — oxygen consumption equals delivery times the fraction extracted. This is the Fick principle, and it is the reason an athlete and a patient with heart failure can be limited by completely different terms.

Expand it fully:

VO₂ = (HR × SV) × [(1.34 × Hgb × SaO₂) + (0.003 × PaO₂)] × E

Six terms. Every one is owned by a different combination of systems, and every one can fail independently.

          OXYGEN DELIVERY, DECOMPOSED — every term and its owner

  VO2 = ( HR × SV ) × [ (1.34 × Hgb × SaO2) + (0.003 × PaO2) ] × E
         └─── CO ──┘   └──────── CaO2, arterial O2 content ───┘  └extraction

  ┌─ HR ────────────── heart rate ───────────────────────────────────┐
  │ SA node intrinsic rate ~100/min, MINUS vagal tone, PLUS beta-1   │
  │ sympathetic drive.  OWNED BY: nervous (autonomic) + cardiac      │
  │ conduction system + adrenal medulla + thyroid.                   │
  │ AMARA: capped by metoprolol. Peak HR 146 against ~176 predicted. │
  └──────────────────────────────────────────────────────────────────┘
  ┌─ SV ────────────── stroke volume ────────────────────────────────┐
  │  PRELOAD     = venous return = blood volume (KIDNEY, ADRENAL,    │
  │                pituitary) × venous tone (NERVOUS) × skeletal     │
  │                muscle pump (MUSCULAR) × respiratory pump         │
  │                (RESPIRATORY) × posture/gravity                   │
  │  AFTERLOAD   = SVR (arterioles, NERVOUS, RAAS, endothelial NO)   │
  │                + aortic stiffness (CONNECTIVE TISSUE, AGE)       │
  │  CONTRACTILITY = sympathetic drive × Ca-handling proteins ×      │
  │                mass of VIABLE myocardium (scar contributes zero) │
  │  COMPLIANCE  = how much volume the ventricle accepts per mm Hg   │
  │                of filling pressure.  AMARA'S LIMITING TERM.      │
  └──────────────────────────────────────────────────────────────────┘
  ┌─ Hgb ───────────── hemoglobin concentration ─────────────────────┐
  │ marrow output (SKELETAL) × erythropoietin (URINARY) × iron,      │
  │ B12, folate absorption (DIGESTIVE) × red cell lifespan           │
  │ (LYMPHATIC/spleen) × absence of hepcidin-driven iron block       │
  │ (IMMUNE).  AMARA: 11.6 g/dL — a 16% cut in every litre pumped.   │
  └──────────────────────────────────────────────────────────────────┘
  ┌─ SaO2 ──────────── arterial saturation ──────────────────────────┐
  │ alveolar ventilation (RESPIRATORY muscle + NERVOUS drive) ×      │
  │ V/Q matching (RESPIRATORY + CARDIOVASCULAR) × diffusion across   │
  │ the 0.3 micrometre blood-gas barrier × inspired PO2              │
  └──────────────────────────────────────────────────────────────────┘
  ┌─ 0.003 × PaO2 ──── dissolved oxygen ─────────────────────────────┐
  │ ~0.3 mL/dL of the ~20 mL/dL total.  Negligible — UNLESS Hgb is   │
  │ near zero or the patient is in a hyperbaric chamber.             │
  └──────────────────────────────────────────────────────────────────┘
  ┌─ E ─────────────── extraction fraction ──────────────────────────┐
  │ capillary density × mitochondrial volume × myoglobin ×           │
  │ 2,3-BPG, pH, temperature (the Bohr effect) × DISTRIBUTION of     │
  │ flow to the tissue that needs it.  OWNED BY: MUSCULAR +          │
  │ CARDIOVASCULAR (microcirculation) + BLOOD.                       │
  │ ~25% at rest whole-body; up to 80-90% in trained muscle.         │
  └──────────────────────────────────────────────────────────────────┘

  THE RULE: functional capacity is a PRODUCT, not a sum.  Cut two terms
  to 80% each and you have 64% of capacity, from two "mild" problems.

(d) Acid–base: two organs, two time constants

Blood pH is defended by a partnership between an organ that acts in minutes and an organ that acts in days, and almost every clinical acid–base problem is an argument about which one moved first.

The relationship you need is conceptual, not algebraic: pH depends on the ratio of bicarbonate to carbon dioxide. Raise the numerator or lower the denominator and pH rises.

• The lung owns the denominator. Alveolar ventilation sets arterial PCO₂, and it can double or halve it within minutes. Central chemoreceptors in the medulla sense CSF pH — which tracks PCO₂ because CO₂ crosses the blood–brain barrier freely and HCO₃⁻ does not — and peripheral chemoreceptors in the carotid and aortic bodies sense PaO₂, PaCO₂, and pH.
• The kidney owns the numerator. It reclaims essentially all filtered bicarbonate in the proximal tubule, and it generates new bicarbonate in the collecting duct by secreting H⁺ onto urinary buffers — phosphate, and ammonia made from glutamine. This takes one to three days to reach full effect and is quantitatively enormous: about 70 mEq of fixed acid per day from an ordinary diet.

Three further participants get left out of the standard account and should not be:

• Bone. Chronic acidosis is buffered by carbonate released from hydroxyapatite. Over years this costs mineral. Amara's chronic bicarbonate of 22 mEq/L, driven by her CKD, is one reason her PTH is rising and her bone density will fall.
• Muscle. The largest protein buffer pool in the body, and the source and consumer of lactate. Lactate is not a waste product; it is a fuel shuttled between fibres and organs.
• Liver. Hepatocytes clear lactate by gluconeogenesis and consume bicarbonate in the urea cycle, so hepatic failure produces both a lactate rise and a loss of an acid-disposal route.

The rule that makes compensation readable: compensation never fully corrects. If the pH is back to exactly 7.40, there is not one disorder compensated — there are two disorders opposing each other. And compensation is always in the same direction as the primary problem's effect on the other term: metabolic acidosis (HCO₃⁻ down) is compensated by hyperventilation (PCO₂ down); respiratory acidosis (PCO₂ up) is compensated by renal bicarbonate retention (HCO₃⁻ up).

(e) The calcium loop: gut, skin, liver, bone, kidney, parathyroid

Ionized calcium is held between 4.5 and 5.3 mg/dL — a range of about 15% — because it sets the threshold for every action potential in the body. Too low and nerves fire spontaneously (tetany, laryngospasm, Chvostek's sign); too high and they will not fire at all (lethargy, constipation, coma).

Sensor: the calcium-sensing receptor on parathyroid chief cells. It is one of the few receptors in the body whose ligand is an ion, and it is inhibitory — falling calcium releases the brake on PTH secretion, which is why PTH responds within seconds.

PTH's three targets, and the fourth organ it recruits:

1. Bone — PTH binds osteoblasts (not osteoclasts), which respond by expressing RANKL, which activates osteoclast precursors. Calcium and phosphate are released together.
2. Kidney — increases distal tubular calcium reabsorption; inhibits proximal phosphate reabsorption, so phosphate is dumped; and activates 1α-hydroxylase.
3. Gut, indirectly — 1α-hydroxylase converts 25-OH vitamin D into calcitriol, which induces the calcium channel TRPV6 and the transport protein calbindin in duodenal enterocytes, raising fractional calcium absorption from ~15% to ~35%.

And the vitamin D pathway is a four-organ relay in its own right: skin (UVB converts 7-dehydrocholesterol to cholecalciferol) → liver (25-hydroxylation) → kidney (1α-hydroxylation) → gut (absorption) → bone (mineralization). Interrupt it anywhere and the endpoint fails.

Amara's version. Her eGFR of 46 means less 1α-hydroxylase, so less calcitriol, so less gut calcium absorption. The same failing kidney excretes phosphate less well, and phosphate retention independently suppresses 1α-hydroxylase and stimulates PTH. Her 25-OH vitamin D was 18 ng/mL when it was first measured (Chapter 6) — a night-shift worker's number — and supplementation has brought it only to 29. The result is a PTH that has risen from 68 pg/mL at Chapter 6 to 88 at Chapter 26 to 96 now, with a normal serum calcium of 9.1 mg/dL: secondary hyperparathyroidism, in which the calcium is normal precisely because PTH is high. She is buying a normal calcium with her skeleton, which is the same currency her mother spent.

(f) The glucose loop: gut, liver, pancreas, muscle, adipose, brain, kidney

The variable: blood glucose, defended asymmetrically. The lower bound is defended ferociously by four counterregulatory hormones because the brain consumes about 120 g of glucose a day and cannot store any. The upper bound is defended by one hormone, insulin, which is why type 2 diabetes exists and hyperglycemic emergencies are common while spontaneous hypoglycemia is rare.

The organs, in order of a meal:

• Gut — SGLT1 in enterocytes absorbs glucose against its gradient using the sodium gradient. Before a single molecule reaches the pancreas, the L and K cells of the intestine release GLP-1 and GIP — the incretins — which prime the β cell. This is why oral glucose produces far more insulin than the same glucose given intravenously.
• Pancreas — β cells sense glucose metabolically: glucose enters via GLUT2, is phosphorylated by glucokinase (the true sensor), raises ATP, closes K-ATP channels, depolarizes the cell, opens voltage-gated Ca²⁺ channels, and releases insulin. α cells do the reciprocal job with glucagon.
• Liver — the only organ that can export glucose in quantity. Insulin switches it from producing glucose (glycogenolysis, gluconeogenesis) to storing it. In insulin resistance, this switch is the first to fail, which is why the fasting glucose rises before the post-meal glucose does.
• Muscle — the largest sink, taking up roughly 80% of a glucose load via GLUT4. Crucially, contraction recruits GLUT4 by an insulin-independent pathway (AMPK), which is the single most important fact in the exercise treatment of diabetes.
• Adipose — stores triglyceride when insulin is high; releases free fatty acids when it is low. In insulin resistance, unrestrained lipolysis floods the liver with FFAs, which worsens hepatic insulin resistance — a feed-forward loop, not a feedback one.
• Brain — obligate consumer; also the sensor for hypoglycemia and the trigger for counterregulation.
• Kidney — reabsorbs about 180 g of filtered glucose per day, 90% of it through SGLT2 in the early proximal tubule, and performs about 20% of the body's gluconeogenesis.

Amara's version. Twenty years of night shift flattened her cortisol rhythm (Chapter 16), visceral adipose released FFAs, her liver and muscle became insulin resistant, her β cells compensated with a fasting insulin of 28 µU/mL, and eventually could not keep up. Eighteen months of walking 4 km a day recruited GLUT4 by contraction; empagliflozin lowered her renal glucose threshold from about 180 mg/dL to roughly 40, so she excretes 60–80 g of glucose a day — 240–320 kcal — in her urine. HbA1c 7.4% → 6.6%.

   THREE CHEMOSTATS — the same architecture, three different currencies

   ┌─ ACID-BASE ────────────────────────────────────────────────────┐
   │  SENSORS   central chemoreceptors (CSF pH) · carotid+aortic    │
   │            bodies (PaO2, PaCO2, pH) · renal tubular cells      │
   │  FAST ARM  LUNG · alveolar ventilation sets PaCO2 · MINUTES    │
   │  SLOW ARM  KIDNEY · H+ secretion onto phosphate and NH3        │
   │            buffers; new HCO3 generated · 1-3 DAYS              │
   │  RESERVE   bone carbonate (years) · muscle protein · liver     │
   │  pH  ∝  HCO3 (kidney, slow)  ÷  PaCO2 (lung, fast)             │
   │  RULE      compensation NEVER fully corrects. pH exactly 7.40  │
   │            with abnormal HCO3 = TWO disorders, not one.        │
   └────────────────────────────────────────────────────────────────┘

   ┌─ CALCIUM ──────────────────────────────────────────────────────┐
   │  SENSOR    parathyroid CaSR — an ion-sensing receptor, seconds │
   │                                                                │
   │   SKIN ──UVB──► cholecalciferol                                │
   │     │                  │                                       │
   │     │             LIVER 25-hydroxylase                         │
   │     │                  │                                       │
   │     │            KIDNEY 1-alpha-hydroxylase ◄── PTH (+)        │
   │     │                  │                        │              │
   │     │             CALCITRIOL                    │              │
   │     │                  ▼                        │              │
   │     └────────────►  GUT: TRPV6 + calbindin       │              │
   │                     Ca absorption 15% → 35%     │              │
   │                          │                      │              │
   │   BONE ◄── PTH → osteoblast RANKL → osteoclast ─┘              │
   │   (the bank: 99% of body Ca, ~1 kg)                            │
   │   KIDNEY: distal Ca REABSORBED, phosphate DUMPED               │
   │   AMARA: eGFR 46 → less calcitriol + phosphate retained        │
   │          → PTH 68 → 88 → 96 pg/mL, Ca still NORMAL at 9.1      │
   └────────────────────────────────────────────────────────────────┘

   ┌─ GLUCOSE ──────────────────────────────────────────────────────┐
   │  SENSOR    pancreatic beta cell glucokinase · hypothalamus     │
   │                                                                │
   │   GUT ──SGLT1──► glucose        GUT ──► GLP-1, GIP (incretins) │
   │            │                              │                    │
   │            ▼                              ▼                    │
   │        PANCREAS  beta cell: INSULIN   alpha cell: GLUCAGON     │
   │            │                              │                    │
   │   ┌────────┼──────────┬─────────┐         │                    │
   │   ▼        ▼          ▼         ▼         ▼                    │
   │ LIVER    MUSCLE    ADIPOSE   KIDNEY     LIVER                  │
   │ store    GLUT4     store TG  SGLT2      glycogenolysis         │
   │ glycogen 80% of    stop      reabsorbs  + gluconeogenesis      │
   │ STOP     the load  lipolysis 180 g/day  ═ THE ONLY EXPORTER    │
   │ output   ▲                   ◄══ empagliflozin                 │
   │          │                                                     │
   │     CONTRACTION recruits GLUT4 WITHOUT insulin (AMPK)          │
   │     ═ why walking 4 km/day lowered Amara's A1c                 │
   │                                                                │
   │  BRAIN  120 g/day, obligate, no store — and the alarm bell     │
   │  ASYMMETRY: 4 hormones defend the floor, 1 defends the ceiling │
   └────────────────────────────────────────────────────────────────┘

33.4 Exercise as the Integration Test

Nothing else you can do voluntarily stresses every system at once. That is why exercise is the healthy-state capstone: it is a controlled, repeatable, whole-body challenge in which every loop in §33.3 must operate simultaneously and correctly, and in which failure of any one of them is immediately visible.

Follow a healthy 24-year-old — Amara's daughter Nia — from the moment she decides to run. The mechanisms invoked below are developed in full in Chapter 32; this section is the integration, not the introduction.

   THE EXERCISE RESPONSE AS A TEMPORAL CASCADE
   time ──────────────────────────────────────────────────────────────►

  t = -10 s   ANTICIPATION.  Motor cortex + hypothalamic locomotor region
  BEFORE      → medulla.  "CENTRAL COMMAND" is feed-FORWARD, not feedback.
  MOVING      Vagal withdrawal.  HR 58 → 82 before the first stride.
              Splanchnic veins constrict; ~500 mL of blood mobilized.
              SYSTEMS: nervous ▸ cardiovascular ▸ digestive

  t = 0       MOTOR RECRUITMENT.  Size principle: small motor units first.
  0-10 s      ATP from stored ATP + phosphocreatine.  No O2 required yet.
              SYSTEMS: nervous ▸ muscular ▸ skeletal (levers, load)

  t = 0-5 s   MUSCLE PUMP.  Rhythmic compression of deep veins.  Venous
              return jumps; preload rises; Frank-Starling raises stroke
              volume within 2-3 beats.  Lymph flow rises 10-30x.
              SYSTEMS: muscular ▸ cardiovascular ▸ lymphatic

  t = 5-30 s  METABOLIC VASODILATION.  Adenosine, K+, H+, CO2, lactate,
              NO, EDHF.  Muscle vascular conductance up to 20x.
              FUNCTIONAL SYMPATHOLYSIS: local metabolites blunt alpha-1
              constriction ONLY in working muscle.
              SYSTEMS: muscular ▸ cardiovascular ▸ nervous

  t = 10-60 s CARDIAC OUTPUT RISES 5 → 25 L/min.  HR 58 → 175; SV 75 →
              125 mL.  MAP rises only ~15-25 mm Hg because conductance
              rises almost as fast as output.  Baroreflex RESETS upward.
              SYSTEMS: cardiovascular ▸ nervous ▸ endocrine

  t = 15-60 s VENTILATION.  Phase I neural (instant, central command +
              joint afferents), phase II metabolic (30-60 s).
              VE 6 → 100+ L/min.  SaO2 barely moves.  Diaphragm and
              intercostals take 10-15% of cardiac output.
              SYSTEMS: respiratory ▸ nervous ▸ muscular

  t = 1-3 min REDISTRIBUTION.  Alpha-1 vasoconstriction cuts splanchnic
              flow ~80% and renal flow ~50%.  GFR held by autoregulation
              + efferent constriction.  Renin, ADH rise; urine output ↓.
              SYSTEMS: cardiovascular ▸ digestive ▸ urinary ▸ endocrine

  t = 3-15min THERMOREGULATION.  Heat production up 15-20x.  Core temp
              +1 to +2 °C.  Skin flow rises (competing with muscle for
              output); sweating up to 1.5-2 L/h.  Plasma volume falls
              10-15% → hemoconcentration → Hgb and CaO2 rise slightly.
              SYSTEMS: integumentary ▸ cardiovascular ▸ urinary ▸ nervous

  t = 10-90min SUBSTRATE MOBILIZATION.  Insulin FALLS (alpha-2 on beta
              cells) while muscle glucose uptake RISES — contraction
              recruits GLUT4 via AMPK, no insulin needed.  Glucagon,
              epinephrine, cortisol, GH rise.  Hepatic glycogenolysis →
              gluconeogenesis.  Adipose lipolysis; FFA oxidation climbs
              as intensity falls.  Muscle IL-6 released as a myokine.
              SYSTEMS: endocrine ▸ digestive/hepatic ▸ muscular ▸
                       adipose ▸ lymphatic-immune

  STOP        RECOVERY.  Vagal reactivation: HR falls 15-25 bpm in the
  0-60 min    first minute (a strong prognostic marker).  EPOC repays
              PCr, lactate, O2 stores, and temperature.  POST-EXERCISE
              HYPOTENSION for 2-8 h.  Muscle glycogen resynthesis is
              maximal in the first 2 h.  Plasma volume OVERSHOOTS within
              24 h (albumin shifted into plasma) — the first chronic
              adaptation, and it begins after ONE session.

  ELEVEN SYSTEMS.  Reproductive is the only one not obligatorily engaged
  — and even it participates, through sex-steroid effects on substrate
  use, and it fails first when energy availability is chronically low.

Three features of that cascade deserve special attention because they are where most students' mental models are wrong.

Feed-forward comes first. Heart rate rises before the first muscle contraction. No feedback loop can explain that, because nothing has happened yet for a receptor to detect. Central command is an anticipatory, open-loop signal — the same copy of the motor plan that goes to the spinal cord also goes to the cardiovascular control centres. Feedback (muscle metaboreflex, mechanoreflex, baroreflex, chemoreflex) then trims the prediction. This is your first honest encounter with a control architecture that is not negative feedback, and §33.10 returns to it.

The baroreflex does not oppose exercise. If the baroreflex simply defended resting pressure, it would fight the rise in cardiac output. Instead the operating point of the reflex is reset upward by central command, so the same reflex now defends a higher pressure — and it still works, which is why you do not faint when you stop suddenly. A reflex whose set point is moved by a higher centre is a recurring pattern (compare fever, §1.5).

Vasodilation and vasoconstriction happen at the same time in the same person. Sympathetic outflow to all vascular beds rises during exercise, including to working muscle. Muscle still dilates, because locally produced metabolites blunt the response of α₁ receptors — functional sympatholysis. Gut and kidney have no such local override, so they constrict. The result is that the body raises total flow and simultaneously redirects it, with one signal and one receptor, using local chemistry as the switch.

33.5 Shock as the Disintegration Test

If exercise is integration working, shock is integration coming apart, and it is the purest demonstration in medicine that the systems are one system.

Shock is not low blood pressure. Shock is inadequate oxygen delivery relative to demand at the cellular level, and blood pressure is only one of the things that can be wrong. A patient can be in shock with a normal blood pressure and can have a low blood pressure without being in shock.

Two equations contain all four categories:

MAP = CO × SVR, and DO₂ = CO × CaO₂, where CO = HR × SV and SV depends on preload, afterload, and contractility.

Each category of shock is the failure of a different term.

   THE FOUR SHOCKS = FOUR FAILED TERMS OF THE SAME TWO EQUATIONS
          MAP = CO × SVR        DO2 = CO × CaO2        CO = HR × SV

  ┌──────────────┬──────────────┬──────────────┬────────────────────┐
  │ HYPOVOLEMIC  │ CARDIOGENIC  │ OBSTRUCTIVE  │ DISTRIBUTIVE       │
  │ PRELOAD fails│ PUMP fails   │ FLOW blocked │ SVR fails (and, in │
  │              │              │ mechanically │ sepsis, EXTRACTION)│
  ├──────────────┼──────────────┼──────────────┼────────────────────┤
  │ hemorrhage   │ large MI     │ tamponade    │ sepsis             │
  │ vomiting     │ arrhythmia   │ tension pneu-│ anaphylaxis        │
  │ burns        │ valve rupture│ mothorax     │ neurogenic (spinal)│
  │ 3rd-spacing  │ myocarditis  │ massive PE   │ adrenal crisis     │
  ├──────────────┼──────────────┼──────────────┼────────────────────┤
  │ CVP    LOW   │ CVP   HIGH   │ CVP   HIGH   │ CVP   LOW / normal │
  │ CO     LOW   │ CO    LOW    │ CO    LOW    │ CO    HIGH         │
  │ SVR    HIGH  │ SVR   HIGH   │ SVR   HIGH   │ SVR   LOW          │
  │ SvO2   LOW   │ SvO2  LOW    │ SvO2  LOW    │ SvO2  HIGH ◄ the   │
  │ skin cold    │ skin cold    │ skin cold    │ skin WARM   tell   │
  │ lactate HIGH │ lactate HIGH │ lactate HIGH │ lactate HIGH       │
  ├──────────────┴──────────────┴──────────────┴────────────────────┤
  │ THE DIAGNOSTIC KEY: three of the four look identical from the   │
  │ bedside — cold, clamped, tachycardic. They are separated by     │
  │ FILLING PRESSURE (low in hypovolemic, high in the other two)    │
  │ and by WHERE the obstruction sits. The fourth inverts every     │
  │ sign, because the failure is in the RESISTANCE term, not flow.  │
  │                                                                 │
  │ In septic shock the lactate is high WHILE SvO2 is high: blood   │
  │ is arriving and the mitochondria cannot use it. That is a       │
  │ failure of the EXTRACTION term, the only term the other three   │
  │ leave intact — and the reason sepsis kills despite good numbers.│
  └─────────────────────────────────────────────────────────────────┘

   THE ORDER OF ORGAN FAILURE, AND WHY IT IS ALWAYS THIS ORDER

   MINUTES   1 BRAIN (function)  autoregulation fails below MAP ~60;
                                 neurons hold ~10 s of ATP. Confusion
                                 and agitation are the FIRST vital sign.
   < 1 HOUR  2 KIDNEY            outer medulla runs at PO2 10-20 mm Hg
                                 NORMALLY (countercurrent costs oxygen);
                                 thick ascending limb is a huge ATP
                                 consumer; renal flow is sacrificed
                                 EARLY by alpha-1. Oliguria = the first
                                 measurable organ injury.
   1-6 HOURS 3 GUT               villus tip sits at the end of a counter-
                                 current hairpin — same trick, same
                                 vulnerability. Barrier fails → bacterial
                                 translocation → the inflammatory
                                 SECOND HIT.
   6-48 h    4 LIVER             dual blood supply protects it at first.
                                 Then: lactate clearance fails (acidosis
                                 worsens), gluconeogenesis fails
                                 (hypoglycemia), clotting factors fail
                                 (coagulopathy).
   12-72 h   5 LUNG              ARDS. Driven by the inflammatory second
                                 hit, NOT by hypoperfusion — which is why
                                 it appears AFTER resuscitation.
   LAST      6 HEART             coronary flow is autoregulated and is
                                 second only to brain in priority. When
                                 it finally fails, the loop closes:
                                 low output → low coronary perfusion →
                                 lower output. POSITIVE FEEDBACK. Death.
   SURVIVE   7 MUSCLE, SKIN, BONE  sacrificed FIRST, fail LAST: low
                                 metabolic rate, large extraction reserve.

   THE RULE:  position in the failure sequence ≈
              (resting O2 extraction ÷ reserve) × (how early the organ is
              sacrificed by the vasoconstriction hierarchy)

The last line of that figure is worth stating on its own, because it is counterintuitive and it is the whole point of studying failure as a network problem:

The organs sacrificed first are not the organs that fail first. Skin and skeletal muscle are shut down within seconds and tolerate it for hours, because their resting oxygen extraction is low and they have enormous reserve. The kidney and gut are sacrificed early and fail early, because they are simultaneously high-priority targets for vasoconstriction and tissues that already run near their extraction ceiling in normal life. The renal medulla's low oxygen tension is not pathological; it is the unavoidable cost of the countercurrent multiplier that lets you concentrate urine. That design choice, made for a good reason, is exactly what makes the kidney the first organ to be injured in shock.

33.6 Amara's Model, Assembled and Run

Here is the whole file. Every system, every direction of causation, every drug.

Read the figure first as a picture, then read the walkthrough, then go back to the figure and find each step in it. That is how an integrated model is meant to be used: as something you navigate, not something you recite.

        AMARA OSEI, 46 — THE COMPLETE MODEL, WITH MEDICATION SITES
        ═══ = the primary vicious cycle    ─── = a contributing arrow
        [1] metoprolol  [2] lisinopril  [3] empagliflozin
        [4] furosemide  [5] atorvastatin  [6] aspirin

   ┌──────────────────────┐        ┌───────────────────────────────┐
   │ NERVOUS              │        │ RESPIRATORY                   │
   │ 20 y night shift →   │───────►│ OSA, AHI was 32/h → nocturnal │
   │ short sleep, flat    │        │ hypoxemia + arousals          │
   │ cortisol rhythm      │◄───────│ → sympathetic surge  ── CPAP  │
   │ SDNN 42→78 ms        │        │ residual AHI 3.1/h            │
   └──────┬───────────────┘        └───────────┬───────────────────┘
          │ sympathetic tone ↑                 │ hypoxemia
          │  [1] blocks beta-1 at BOTH         │ → EPO signal, and
          │      the SA node and the           │ → pulmonary vasoconstriction
          │      juxtaglomerular cell          │
          ▼                                    ▼
   ╔══════════════════════════════════════════════════════════════════╗
   ║ HEART       EF 48%   E/e' 15   NT-proBNP 410   NYHA II           ║
   ║ Diastolic stiffness from: LVH (afterload) + interstitial         ║
   ║ collagen (angiotensin II, aldosterone) + inferolateral SCAR      ║
   ║ (the NSTEMI — scar is collagen, and collagen does not contract   ║
   ║  or relax; Chapter 4)                                            ║
   ║ [1] rate ↓ 104→62 → longer diastole, lower myocardial O2 demand  ║
   ║ [2] less angiotensin II → less fibrosis, lower afterload         ║
   ╚══╦═══════════════════════════════════════════════════════════╦═══╝
      ║ high filling pressure                    low forward flow ║
      ║ transmitted backwards                    + renal venous   ║
      ▼                                          congestion       ▼
   ┌──────────────────────┐                  ┌───────────────────────┐
   │ LUNG (again)         │                  │ KIDNEY  eGFR 46       │
   │ pulmonary venous     │                  │ ACR 165 mg/g          │
   │ pressure ↑ → inter-  │                  │ Senses LOW PERFUSION  │
   │ stitial fluid → J-   │                  │ (not low volume)      │
   │ receptors → DYSPNEA  │                  │ → RENIN ↑             │
   │ [4] lowers volume    │                  │ [3] restores tubulo-  │
   │     → relieves this  │                  │     glomerular        │
   └──────────────────────┘                  │     feedback          │
                                             │ [4] blocks NKCC2 —    │
   ┌──────────────────────┐                  │     and RAISES renin  │
   │ LIVER (digestive)    │◄─────────────────┤ [2] dilates efferent  │
   │ angiotensinogen —    │  renin cleaves   │     arteriole         │
   │ the RAAS substrate   │  it here         └───────┬───────────────┘
   │ [5] HMG-CoA reduc-   │                          │ ANG I
   │     tase → LDL 62    │                          ▼
   └──────────────────────┘                  ┌───────────────────────┐
                                             │ LUNG endothelial ACE  │
   ┌──────────────────────┐                  │ ANG I ──► ANG II      │
   │ ADRENAL CORTEX       │◄─────────────────┤ [2] BLOCKED HERE      │
   │ aldosterone → ENaC   │  ANG II          └───────────────────────┘
   │ → Na+ retained,      │
   │ K+ and H+ lost       │──────► PLASMA VOLUME ↑ ──► PRELOAD ↑ ═══► HEART
   └──────────────────────┘                              (loop closes)

   ┌─────────────────────────────────────────────────────────────────┐
   │ THE OTHER SYSTEMS, ALL FEEDING THE SAME LOOP                    │
   │                                                                 │
   │ BLOOD   Hgb 11.6 (renal EPO ↓, iron TSAT 19%, hepcidin) →       │
   │         every litre of output carries 16% less O2 ─► exercise   │
   │ BONE    eGFR 46 → less calcitriol; phosphate retained → PTH 96  │
   │         → bone resorption ─► Adwoa's diagnosis, one generation  │
   │         earlier                                                 │
   │ MUSCLE  deconditioning → low capillary + mitochondrial density  │
   │         → low extraction.  REVERSIBLE — and it was reversed.    │
   │ ADIPOSE visceral fat → FFA to liver + adipokines → insulin      │
   │         resistance ─► the ORIGIN of the whole picture           │
   │ IMMUNE  plaque is an inflammatory lesion; [6] blocks platelet   │
   │         COX-1 so a ruptured plaque is less likely to occlude    │
   │ VESSEL  PWV 9.6 m/s: a stiff aorta returns the reflected        │
   │         pressure wave in SYSTOLE, raising afterload on an       │
   │         already stiff ventricle                                 │
   │ REPRODUCTIVE  perimenopause → falling oestrogen → less          │
   │         endothelial NO → stiffer vessels (Chapter 27)           │
   │                                                                 │
   │ POTASSIUM, held at 4.6 mEq/L by FOUR opposing forces:           │
   │   [4] furosemide drives it OUT   [2] lisinopril holds it IN     │
   │   eGFR 46 excretes it poorly     muscle buffers ~3,000 mEq      │
   │ Change ANY one and the number moves. The normal value is the    │
   │ most actively defended number on her panel.                     │
   └─────────────────────────────────────────────────────────────────┘

Running the model

Where it started. Not the heart. Visceral adiposity and twenty years of circadian disruption produced insulin resistance and hypertension (Chapters 12, 16, 24). Hypertension plus insulin resistance plus dyslipidemia produced coronary plaque; plaque rupture produced the NSTEMI of Chapter 18. The infarct left a region of the inferolateral wall as scar — collagen, not muscle, and collagen neither contracts nor relaxes (Chapter 4). Meanwhile the same hypertension had already produced concentric hypertrophy, and hypertrophied myocardium with interstitial collagen is a stiff myocardium.

Why she is symptomatic despite an EF of 48%. Ejection fraction reports what fraction of the end-diastolic volume leaves the ventricle. It says nothing about how much volume got in, or at what pressure. Amara's problem is on the filling side: an E/e′ of 15 means she reaches adequate end-diastolic volume only by running a high left atrial pressure. At rest that is tolerable. On exertion, when heart rate rises and diastole shortens, filling gets worse precisely when demand is highest, so left atrial pressure climbs, pulmonary capillary pressure follows, fluid enters the interstitium, J-receptors fire, and she is breathless. This is heart failure with a nearly normal ejection fraction, and it is the commonest form of heart failure in women over forty-five.

Note where 48% falls. The guideline categories put heart failure with preserved ejection fraction at 50% or above and label 41–49% mildly reduced (Chapter 18). Her file is written as HFpEF because her physiology is diastolic, but she sits on the line by two percentage points, and a repeat echocardiogram read by a different sonographer could move her across it without anything changing inside her chest. The category is a threshold drawn across a continuum by people who needed a threshold. It is the same low-resolution problem as the eleven organ systems, one level down, and it is worth noticing that the number that names her disease is the least informative number on her chart.

Why the kidney keeps getting worse. Three mechanisms in parallel, none sufficient alone: reduced forward flow; raised renal venous pressure, which lowers the pressure gradient across the kidney more effectively than a fall in arterial pressure does; and years of glomerular capillary hypertension from diabetes and hypertension. The RAAS activation that follows is a rational response to what the kidney senses and a disaster for what the heart needs.

The closing loop, in one sentence. A stiff ventricle raises filling pressure, which congests the renal veins and lowers renal perfusion, which raises renin, which raises angiotensin II and aldosterone, which retain sodium and raise plasma volume and afterload and lay down myocardial collagen, which makes the ventricle stiffer.

That is a positive feedback loop built out of negative feedback components. Every element is behaving correctly with respect to its own variable. The kidney is defending perfusion. The adrenal is defending volume. The sympathetic nervous system is defending pressure. Each is right; the system is wrong. This is the single most important pattern in chronic disease, and once you see it you will see it everywhere.

Where the six drugs interrupt it. Metoprolol at two points (sinoatrial node, granular cell). Lisinopril at the pulmonary ACE step, with effects downstream at the arteriole, the adrenal, the efferent arteriole, and the cardiac fibroblast. Empagliflozin at the proximal tubule, restoring tubuloglomerular feedback and lowering intraglomerular pressure while also removing 60–80 g of glucose daily. Furosemide at NKCC2, lowering volume and preload — symptomatically the most powerful of the six and the only one that activates the loop it is being used to treat. Atorvastatin in the hepatocyte, protecting supply by stabilizing plaque. Aspirin on platelet COX-1, making the next rupture less likely to occlude.

Five of the six act on organs other than the one whose disease they are named for. That is not a curiosity. It is the practical content of this entire chapter.

33.7 The Reasoning Skill, Made Explicit

Everything so far has been demonstration. This section is the method itself, stated as five steps you can apply to a patient you have never seen, in a system you have half forgotten.

It is deliberately mechanical. Expert clinicians do not consciously run these steps — but they did once, and this is what they were running.

              THE CROSS-SYSTEM REASONING METHOD

   ┌─────────────────────────────────────────────────────────────────┐
   │ STEP 1 · NAME THE VARIABLE THAT IS OFF                          │
   │  Not the diagnosis. Not the organ. The regulated VARIABLE.      │
   │  Use the list of nine (Fig 28.2). If several are off, ask which │
   │  is off FIRST in time — the others may be its consequences.     │
   └────────────────────────────┬────────────────────────────────────┘
                                ▼
   ┌─────────────────────────────────────────────────────────────────┐
   │ STEP 2 · LIST EVERY SYSTEM THAT DEFENDS IT                      │
   │  Six to eleven of them, every time. Read them off Figure 33.2.  │
   │  You now have your complete differential — the failure MUST be  │
   │  in one of these, because nothing else touches the variable.    │
   └────────────────────────────┬────────────────────────────────────┘
                                ▼
   ┌─────────────────────────────────────────────────────────────────┐
   │ STEP 3 · SEPARATE THE PRIMARY FAILURE FROM THE COMPENSATIONS    │
   │  TEST: does this finding move the variable TOWARD normal or     │
   │  AWAY from it?                                                  │
   │     toward  → compensation. Do not treat it. It is helping.     │
   │     away    → the primary problem, or a second one.             │
   │  Most abnormal numbers on any chart are compensations.          │
   └────────────────────────────┬────────────────────────────────────┘
                                ▼
   ┌─────────────────────────────────────────────────────────────────┐
   │ STEP 4 · ASK WHAT EACH COMPENSATION COSTS                       │
   │  Every compensation spends something: oxygen, substrate,         │
   │  another variable, or a tissue.  Tachycardia buys output and    │
   │  spends myocardial O2 and diastolic filling time.               │
   │  Hyperventilation buys pH and spends respiratory muscle work.   │
   │  Vasoconstriction buys pressure and spends the gut and kidney.  │
   │  The cost is where the NEXT problem will appear.                │
   └────────────────────────────┬────────────────────────────────────┘
                                ▼
   ┌─────────────────────────────────────────────────────────────────┐
   │ STEP 5 · LOOK FOR THE LOOP WHERE A COMPENSATION FEEDS THE       │
   │          ORIGINAL PROBLEM                                        │
   │  If you find one, you have found why the patient is             │
   │  deteriorating rather than stabilizing — and you have found     │
   │  the intervention, because breaking the loop is usually more    │
   │  effective than treating either end of it.                      │
   │  If you find none, the patient will probably compensate and     │
   │  stabilize. Say so; it is a prediction, and it can be checked.  │
   └─────────────────────────────────────────────────────────────────┘

   THE TWO QUESTIONS THAT DRIVE ALL FIVE STEPS
     FORWARDS:  given this mechanism, what happens next?
     BACKWARDS: for this finding to exist, what must be true upstream?

Worked case A — Toby's roommate: a nineteen-year-old with diabetic ketoacidosis

Presentation. Nineteen, new-onset type 1 diabetes. Four days of thirst, polyuria, weight loss; twelve hours of vomiting. HR 128, BP 96/58, RR 34 and deep, breath smells of acetone. Glucose 512 mg/dL. pH 7.08, PaCO₂ 15 mm Hg, HCO₃⁻ 6 mEq/L. Na⁺ 128 mEq/L, K⁺ 5.4 mEq/L, Cl⁻ 96, creatinine 1.4 mg/dL. Anion gap 26. Ketones strongly positive.

Step 1 — the variable. Several are off, so order them in time. Glucose moved first; insulin deficiency made it rise. pH moved second, because unrestrained lipolysis delivers fatty acids to a liver that, with no insulin and high glucagon, converts them to acetoacetate and β-hydroxybutyrate — strong acids. Osmolality and sodium moved third, through osmotic diuresis. Potassium is the trap and is discussed below.

Step 2 — who defends each. Glucose: pancreas, liver, muscle, adipose, gut, kidney, brain. pH: lung, kidney, blood buffers, bone, muscle, liver. Osmolality: kidney, posterior pituitary, hypothalamus, gut. The primary failure is in exactly one node — the pancreatic β cell — and everything else on the chart is downstream.

Step 3 — primary versus compensation.

Step 4 — the costs. Kussmaul breathing buys about 0.3 pH units and spends respiratory muscle work; when the patient tires, PaCO₂ rises and pH collapses, so a rising PaCO₂ in DKA is an ominous sign, not an improving one. The osmotic diuresis protects against extreme hyperglycemia and spends volume, potassium, magnesium, and phosphate.

Step 5 — the loop. Vomiting and osmotic diuresis reduce volume; reduced volume reduces renal perfusion; reduced renal perfusion reduces glucose and ketone clearance; higher glucose worsens the osmotic diuresis. That is the loop, and it is why fluid, not insulin, is the first treatment. It also predicts the classic iatrogenic disaster: give insulin first and potassium moves back into cells with glucose, and a serum potassium of 5.4 with a depleted total body store becomes 2.6 within two hours — arrhythmia and death. Insulin is withheld until potassium is known to be above about 3.3 mEq/L and replacement has begun.

Five steps, one unfamiliar patient, and the two most important management decisions fall out of the physiology rather than from a memorized protocol.

Worked case B — a sixty-eight-year-old with COPD and cor pulmonale

Presentation. Fifty pack-years. Chronic dyspnea, ankle swelling, distended neck veins, pulsatile liver. pH 7.36, PaCO₂ 62 mm Hg, HCO₃⁻ 34 mEq/L, PaO₂ 54 mm Hg, SaO₂ 87% on air. Hematocrit 57%. Echo: dilated, poorly contracting right ventricle; estimated pulmonary artery systolic pressure 58 mm Hg; normal left ventricle.

Step 1. Two variables: arterial oxygen (primary) and pH (secondary, and currently near-normal).

Step 2. Oxygen is defended by lung, heart, vessels, blood, marrow, kidney (EPO), gut (iron), and the nervous system (drive). pH by lung and kidney.

Step 3. The primary failure is alveolar: destroyed alveolar surface plus airflow obstruction produce V/Q mismatch and hypoventilation. Everything else is compensation. The HCO₃⁻ of 34 is renal compensation for chronic respiratory acidosis — a normal kidney doing its job over weeks, and the reason the pH is 7.36 rather than 7.15. The hematocrit of 57% is secondary polycythemia: chronic hypoxemia raises renal erythropoietin, the marrow responds, and arterial oxygen content is defended even though saturation cannot be.

Step 4 — the costs, and this is where the case becomes interesting.

• Polycythemia buys oxygen content and spends viscosity. Blood viscosity rises steeply above a hematocrit of about 55%, raising resistance everywhere, including in the pulmonary circuit, and raising thrombotic risk.
• Renal bicarbonate retention buys pH and spends respiratory drive: with CSF pH normalized, the central chemoreceptors stop contributing, and the patient's ventilation becomes more dependent on hypoxic drive from the carotid bodies. This is the real, narrow, and often overstated basis for caution with high-flow oxygen in chronic CO₂ retainers — though the larger mechanism is loss of hypoxic pulmonary vasoconstriction worsening V/Q matching, plus the Haldane effect.
• Hypoxic pulmonary vasoconstriction is the crucial one. In a healthy lung it is elegant: poorly ventilated alveoli constrict their own arterioles, diverting blood to better ventilated regions. It is the only vascular bed in the body that constricts in response to low oxygen, and it exists to match perfusion to ventilation. When hypoxemia is global rather than regional, the whole pulmonary bed constricts, and the mechanism becomes a disease: pulmonary hypertension.

Step 5 — the loop. Global hypoxemia → generalized hypoxic pulmonary vasoconstriction → pulmonary vascular resistance up → right ventricular afterload up → RV dilates and fails → systemic venous congestion (the neck veins, the liver, the ankles) → and, because a dilated RV bows the interventricular septum leftward, reduced LV filling → lower cardiac output → worse tissue oxygen delivery. Cor pulmonale is right heart failure caused by the lung, and the mechanism is a normal reflex applied to an abnormal situation.

The intervention follows directly: long-term oxygen therapy, targeting SaO₂ 88–92%, is one of only two treatments shown to prolong life in COPD. It works by relieving hypoxic pulmonary vasoconstriction — that is, by breaking the loop rather than by treating either the lung or the heart.

Worked case C — a twenty-four-year-old with hemorrhagic shock

Presentation. Motorcycle collision, 40 minutes ago. BP 82/54, HR 138, RR 30, cool mottled skin, capillary refill 4 s, confused. Hemoglobin 12.4 g/dL — normal. Lactate 6.8 mmol/L, base deficit −9. Temperature 34.8 °C. INR 1.6. Estimated blood loss 1,800 mL.

Step 1. Mean arterial pressure ≈ 63 mm Hg, and — more importantly — oxygen delivery.

Step 2. MAP is defended by all ten systems in Figure 33.2. Oxygen delivery is defended by heart, vessels, blood, marrow, lung, kidney, gut.

Step 3. Primary failure: the preload term. Not the pump, not resistance, not oxygen content. The single most instructive number is the normal hemoglobin: whole blood was lost, so the concentration is unchanged until interstitial fluid refills the vascular space over the following hours. A normal hemoglobin in acute hemorrhage is expected and reassures nobody. Compensations: tachycardia (buys output), vasoconstriction (buys pressure, explaining cool mottled skin and capillary refill), tachypnea (buys pH), and transcapillary refill (buys volume).

Step 4 — the costs, which here are lethal.

• Vasoconstriction buys blood pressure and spends the kidney and gut — the first organs to be injured (§33.5).
• Anaerobic metabolism buys ATP and spends pH — lactate 6.8, base deficit −9.
• Every unit of cold crystalloid or stored blood buys volume and spends temperature and clotting factor concentration.

Step 5 — the loop, which has a name. The lethal triad: hypothermia, acidosis, and coagulopathy, each worsening the other two. Clotting enzymes are enzymes, and enzyme kinetics are temperature- and pH-dependent; at 34 °C and pH 7.2, the coagulation cascade runs at a fraction of normal speed regardless of how many factors are present. Impaired clotting means continued bleeding, which means more hypoperfusion, which means more acidosis, and more transfusion, which means more cold and more dilution.

The management follows from the loop, and it is precisely why modern trauma care looks the way it does: stop the bleeding first, keep the patient warm, avoid large-volume crystalloid, transfuse plasma and platelets alongside red cells rather than after them, and accept a lower-than-normal blood pressure until surgical control is achieved. Every one of those is an attack on the loop rather than on a number.

33.8 What This Book Has Been Arguing

Three threads have run through every chapter. It is worth saying plainly, now that the evidence is in, that they were never three claims. They were one claim, examined from three directions.

Thread 1 — Structure determines function. You met it as the thick left ventricle and the vast thin alveolar surface. You saw it become predictive: the countercurrent architecture of the renal medulla predicts both the ability to concentrate urine and the medulla's vulnerability to ischemia (§33.5). The developmental fact that cardiac myocytes leave the cell cycle predicts that Amara's infarct becomes collagen, and collagen predicts a stiff ventricle, and a stiff ventricle predicts breathlessness on exertion with a nearly normal ejection fraction. Structure was never decoration; it was the constraint that generated everything downstream.

Thread 2 — Homeostasis is the master concept. You met it as nine variables and a four-box control loop. Section 28.2 showed what the table actually claims: those nine variables are defended by six to ten organ systems apiece, layered in time from milliseconds to weeks, each layer buying time for the next. And §33.7 gave the clinical payoff — that most abnormal numbers are defences rather than defects, and that the danger usually lies in what the defence costs.

Thread 3 — The body is integrated. You met it as Amara's cool, pale skin: an integumentary sign produced by a nervous-system decision about a cardiovascular emergency. Everything since has been the same observation at higher resolution. The matrix in §33.1 puts a number on it — 92% of ordered system pairs carry a direct regulatory link. The RAAS cannot be described without five organs. Oxygen delivery cannot be decomposed without eight. Shock is not a cardiovascular event; it is a network collapsing in a fixed order determined by who was sacrificed to protect whom.

The single claim underneath all three: the body is a set of regulated variables defended by loops that run through multiple organs, and structure is what makes each loop physically possible. Organ systems are how we file it. Loops are how it works.

That is the whole argument. Twenty-seven chapters were the evidence.

33.9 Where You Go Next

An honest map. Each of these courses will assume you already own specific things from this book, and will not re-teach them.

Pathophysiology. The most direct sequel and the course this book was shaped to feed. It assumes Chapter 1's feedback architecture, Chapter 4's tissue responses to injury (the fact that scar is not the original tissue is the organizing idea of half the course), Chapter 20's inflammatory mechanisms, and every loop in §33.3. It will move fast through mechanism and spend its time on how mechanisms fail. If you own §33.7's five steps, most of pathophysiology is applying them to one organ at a time.

Pharmacology. Assumes receptors and second messengers (Chapter 16), membrane transport and channels (Chapter 11), autonomic receptor subtypes and their distribution (Chapter 13), hepatic metabolism (Chapter 23) and renal excretion (Chapter 26), and the RAAS and adrenergic loops in §33.3. Every drug class you learn will be a named point on a loop you already know. If you cannot draw the loop, you will memorize the drug; if you can, you will derive it — including its side effects, which are simply the loop's other branches.

Biomechanics and kinesiology. Assumes joint structure and lever classes (Chapter 7), excitation–contraction coupling, the length–tension and force–velocity relationships (Chapter 9), muscle architecture and moment arms (Chapter 10), motor unit recruitment and the size principle (Chapter 11), proprioception and reflex modulation (Chapter 13), and bone's response to load (Chapter 6). Add §33.4 for the cardiopulmonary limits on sustained work.

The MCAT and similar exams. They test §33.3 almost directly. Expect pressure–volume loops, the Fick principle, acid–base disturbances requiring compensation calculations, countercurrent multiplication, and drug effects on named steps of the RAAS and the baroreflex. The Level 4 questions at the end of every chapter of this book are written at that level deliberately. Weight Chapters 3, 10, 14, 16, 17, 19, 22, 25, 27, and this one — see Learning Paths, Path 2.

Clinical rotations. Assumes you can convert a vital sign into a mechanism in real time. The specific habits that transfer: reading an abnormal number as a compensation before reading it as a defect; knowing which organ fails first and why (§33.5); knowing that lactate trajectory beats lactate value; knowing that a normal number can be actively and precariously defended (Amara's potassium, the crush-injury potassium in §33.2). Nobody will ask you to recite the eleven organ systems. Someone will ask you why the creatinine went up when you increased the diuretic, and you now know.

Nursing specifically. Chapter 31 and §33.2 are the two highest-yield passages in the book for you, with Chapter 26 immediately behind them. Most hospitalized patients are fluid, electrolyte, and acid–base problems wearing a diagnosis, and the four-moves framework — shift, buffer, excrete, re-set demand — covers most of what you will do at the bedside.

33.10 Advanced Topic · The Limits of the Integrated Model

This book has argued for a particular picture: defended variables, negative feedback loops, organs as nodes. That picture is powerful and it is incomplete. Four honest qualifications.

1 · Allostasis: regulation by prediction, not by error. Homeostasis is error-correcting — something deviates, a receptor detects it, a response opposes it. But you saw in §33.4 that heart rate rises before exercise begins, which no error-correcting loop can explain. The alternative framework, allostasis, proposes that the brain regulates the body largely by prediction: it anticipates demand from context and prior experience and adjusts the body in advance, and it treats set points as adjustable outputs rather than fixed inputs.

The distinction matters clinically because it changes what a chronic abnormality means. Under homeostasis, Amara's years of high blood pressure were a broken loop. Under allostasis, they were the correct output of a control system responding to sustained demands — night shift, short sleep, chronic stress, high sodium — a system that shifted the defended level upward because the anticipated requirement was higher. The cost of running at that new level for years is called allostatic load, and it is measurable: left ventricular mass, glomerular pressure, arterial stiffness, HbA1c. This is a better account of most chronic disease than "a feedback loop broke," and physiology is still working out how to reconcile the two frameworks.

A related distinction worth having: a set point is a defended target; a settling point is simply where opposing processes happen to balance, with nothing defending it. Body weight in adults may be closer to a settling point than a set point, which would explain a great deal about why weight is so hard to change and why it is regained.

2 · Individual variation, and what a reference range actually is. A reference range is the central 95% of values measured in a reference population. It follows that one healthy person in twenty falls outside any given range by definition, and that a panel of twenty independent tests has a better-than-even chance of producing at least one "abnormal" result in a completely well person.

More importantly, ranges are population statistics applied to individuals. A person's own biological variation is often far narrower than the population range. A creatinine of 1.0 mg/dL is normal; if that person's creatinine has been 0.6 for a decade, it represents a 40% loss of filtration and is not normal for them. The serial value beats the reference range, which is the same lesson as the lactate trajectory and the crush-injury potassium. The direction physiology is moving is toward n-of-1 baselines, and the technology to do it routinely now exists.

3 · Complexity and emergent behavior. Some of what a body does cannot be derived from its loops even in principle.

• Heart rate variability is not noise. The interval between heartbeats fluctuates in a complex, partly fractal pattern, and losing that complexity — as Amara's SDNN of 42 ms did — predicts mortality better than many conventional measures. A healthy system is variable; a dying one becomes regular. No simple negative feedback loop predicts that ordering.
• Critical transitions. Some deteriorations are not gradual. Systems near a tipping point show characteristic early warning signs — rising variance, slower recovery from small perturbations — and then change state abruptly. Decompensated heart failure, septic collapse, and status epilepticus may all be transitions of this kind. The clinical implication is uncomfortable: a patient can be genuinely stable and genuinely near collapse at the same time.
• Network position beats node severity. As §33.6 showed, what predicts disability is often not how badly a node is damaged but where it sits. Two mildly damaged coupled nodes outperform — in the wrong direction — one severely damaged isolated node.

4 · What physiology still does not explain. A short and genuinely open list.

• What sets long-term blood pressure. Pressure natriuresis is the mechanism with infinite gain, but why the kidney's pressure–natriuresis curve sits where it does in a given person, and what moves it in essential hypertension, is not settled.
• Why some people with identical risk factors get HFpEF and others do not. Amara's trajectory is common but not predictable at the individual level, and there is no biomarker that identifies it early.
• Aging. There is no accepted mechanistic theory that explains why maximum function declines at the rate it does across every system roughly in parallel. There are at least a dozen candidate mechanisms and no accounting that adds up.
• The genotype-to-physiology gap. Polygenic scores for blood pressure, body mass, and coronary disease explain a modest fraction of variance. We cannot compute a person's physiology from their genome and are not close.
• The microbiome. Roughly as many bacterial cells as human ones, demonstrably influencing metabolism, immunity, and the nervous system — and it is not one of the eleven systems, has no chapter in this book, and is not yet mechanistically integrated into clinical physiology.
• Why exercise works. Exercise improves nearly every outcome in nearly every disease. The list of mechanisms is long and each is real. Why the combination is so much more effective than any single pathway would predict is not understood.
• Consciousness. The nervous system's outputs are describable in the terms of Chapters 11 to 13. Its most obvious property is not.

None of this weakens the model you have built. A map is not discredited by having edges. But the honest position at the end of a first course in physiology is that you now hold the best available working model of the human body, that it is astonishingly good, and that it is not finished — and that the people who will finish it will be readers of books like this one.

Chapter Summary

§33.1 The eleven-organ-system partition is a map of anatomical contiguity, built from dissection, applied to a body whose actual organization is functional coupling. It hides organs with two citizenships, cross-boundary regulation (101 of 110 ordered system pairs carry a direct influence), variables that no single system owns, the fixed order in which organs fail, and the fact that every useful drug acts in more than one system. What replaces it is not "everything affects everything" but a short list of defended variables and a finite set of loops. Organs are nodes; the loops are the physiology.

§33.2 Each of the nine variables from §1.5 is defended by six to ten organ systems acting primarily. The defenders are layered in time — physicochemical buffering, neural reflexes, fast and slow endocrine responses, fluid shifts, renal excretion, and synthesis — and the fast layers are weak while the strong layers are slow. Any abnormal value is therefore the residue left after every defender has already acted, and any normal value may be actively and expensively defended. Every response is one of four moves: shift, buffer, produce or excrete, or re-set demand.

§33.3 Six loops carry most of clinical physiology. The RAAS runs through kidney, liver, lung, adrenal, and back to kidney and heart, defends effective circulating volume rather than blood volume, and doubles as a fibrosis pathway. The baroreflex runs vessel → cranial nerves IX and X → nucleus of the solitary tract → heart and vessels in under a second, and resets, which is why it cannot set long-term pressure — renal pressure natriuresis does. Oxygen delivery decomposes into heart rate, stroke volume, hemoglobin, saturation, dissolved oxygen, and extraction, each owned by a different combination of systems, combined multiplicatively. Acid–base is a two-organ partnership with a minutes-long arm and a days-long arm, backstopped by bone, muscle, and liver. The calcium loop relays through skin, liver, kidney, gut, bone, and parathyroid. The glucose loop runs through gut, pancreas, liver, muscle, adipose, brain, and kidney, and is defended asymmetrically — four hormones for the floor, one for the ceiling.

§33.4 Exercise engages every system in a fixed temporal order: anticipatory central command before movement, motor recruitment, muscle pump, metabolic vasodilation with functional sympatholysis, a fivefold rise in cardiac output, the two-phase ventilatory response, splanchnic and renal vasoconstriction, thermoregulation with a 10–15% fall in plasma volume, hormonal substrate mobilization with insulin-independent GLUT4 recruitment, and a recovery whose first chronic adaptation — plasma volume expansion — is complete within a day.

§33.5 The four shocks are failures of different terms of MAP = CO × SVR and DO₂ = CO × CaO₂: preload (hypovolemic), pump (cardiogenic), mechanical obstruction (obstructive), and resistance plus extraction (distributive). Three look identical at the bedside and are separated by filling pressure; the fourth inverts every sign, and a high mixed venous saturation with a high lactate is its signature. Organs fail in a fixed order — brain function, kidney, gut, liver, lung, heart — set by resting extraction relative to reserve and by position in the vasoconstriction hierarchy. Organs sacrificed first are not the organs that fail first.

§33.6 Amara's picture is one loop with several impaired nodes: a stiff ventricle raises filling pressure, which congests renal veins and reduces renal perfusion, which raises renin, which raises angiotensin II and aldosterone, which retain sodium and raise preload and afterload and lay down myocardial collagen, which stiffens the ventricle further. It is a positive feedback loop built entirely from correctly functioning negative feedback components. Her six drugs interrupt it at eight identifiable points, and five of the six act on an organ other than the one their indication names.

§33.7 The method: name the variable, list every system that defends it, separate primary failure from compensation by asking whether each finding moves the variable toward or away from normal, ask what each compensation costs, and look for the loop in which a compensation feeds the original problem. Worked forward through diabetic ketoacidosis (fluid before insulin; the potassium trap), COPD with cor pulmonale (hypoxic pulmonary vasoconstriction as a normal reflex applied globally), and hemorrhagic shock (the lethal triad, and why a normal hemoglobin reassures nobody).

§33.8 The three threads were one claim: the body is a set of regulated variables defended by loops that run through multiple organs, and structure is what makes each loop physically possible.

§33.9 Pathophysiology assumes §33.3 and Chapters 1, 4, and 20. Pharmacology assumes receptors, transporters, autonomic subtypes, and the loops, because drug classes are named points on them. Biomechanics assumes Chapters 6 through 10. The MCAT tests §33.3 nearly directly. Rotations assume you can read a vital sign as a compensation in real time.

§33.10 The model has edges: allostasis and predictive regulation, settling points versus set points, reference ranges as population statistics misapplied to individuals, loss of complexity as a mortality signal, critical transitions, and a genuine list of things physiology cannot yet explain — including what sets long-term blood pressure, why exercise works as well as it does, and consciousness.

Case File 33 · Resolution

Question 1 — Trace one complete causal loop that begins and ends at her heart, through at least four other organ systems, and name every point where a drug intervenes.

Start at the ventricle at end-diastole.

Heart → Kidney (urinary). Amara's stiff left ventricle fills only by running a high left atrial pressure (E/e′ 15). That pressure is transmitted backward through the pulmonary circulation to the right heart and into the systemic veins, so renal venous pressure rises. At the same time forward output is limited. Renal perfusion is a gradient — arterial pressure minus venous pressure — so a congested renal vein reduces renal perfusion just as effectively as a low arterial pressure, and in heart failure it is often the larger term. Her juxtaglomerular granular cells sense reduced afferent arteriolar stretch and, via the macula densa, reduced sodium delivery. Renin is released.

Kidney → Liver (digestive). Renin has no effect on its own; it is a protease. Its substrate, angiotensinogen, is synthesized continuously by hepatocytes and circulates in excess. Renin cleaves it to angiotensin I.

Liver → Lung (respiratory). Angiotensin I is inert until angiotensin-converting enzyme, anchored on pulmonary capillary endothelium, removes two residues to make angiotensin II. Every litre of her cardiac output passes this enzyme every minute.

Lung → Adrenal cortex (endocrine). Angiotensin II stimulates the zona glomerulosa to release aldosterone, and simultaneously acts directly on arterioles (raising systemic resistance and therefore her afterload), on the efferent arteriole (preserving GFR), on the posterior pituitary (ADH, water retention), on the subfornical organ (thirst), and on cardiac and renal fibroblasts.

Adrenal → Kidney again. Aldosterone increases ENaC and Na⁺/K⁺-ATPase in collecting duct principal cells over one to three hours. Sodium is retained; potassium and hydrogen ion are lost.

Kidney → back to the Heart. Sodium retention expands plasma volume, which raises venous return and preload. In a normal ventricle that would raise stroke volume. In hers, the Frank–Starling relationship is flat, so the extra volume raises pressure without raising output — and the raised pressure returns to the kidney, closing the loop. Meanwhile angiotensin II and aldosterone are laying interstitial collagen into her myocardium, which makes the ventricle stiffer still, which raises the filling pressure required next time.

Nervous system runs alongside the whole loop. Reduced effective arterial filling unloads the carotid and aortic baroreceptors; sympathetic outflow rises; renal sympathetic nerves act on β₁ receptors of the same granular cells to release more renin, while α₁ constriction raises afterload and higher heart rate shortens diastole — which, in a ventricle that fills poorly, is exactly the wrong direction.

Systems traversed, not counting the heart: urinary, digestive/hepatic, respiratory, endocrine (adrenal and pituitary), nervous, and connective tissue. Six.

Where the drugs intervene — eight points:

Question 2 — Why do two mild abnormalities in coupled systems produce more disability than one severe abnormality in an isolated system?

Four reasons, and the first is arithmetic.

(a) Capacity is a product, not a sum. From §33.3(c), VO₂ = CO × CaO₂ × E. Compare Amara with a predicted sedentary 46-year-old woman of her size:

0.70 × 0.83 × 1.03 = 0.60. Two "mild" deficits — a cardiac output at 70% of predicted, an oxygen content at 83% — multiply to a functional capacity of 60%, rescued slightly by trained extraction. Neither term alone would be disabling. Their product is.

(b) Each system's compensation for the other's failure is the thing the other cannot tolerate. The normal renal response to a low-output heart is to retain sodium, raising preload. That works when the Frank–Starling curve is steep. Amara's is flat, so the retained volume becomes pulmonary congestion rather than stroke volume. Conversely, the normal cardiac response to a hypoperfused kidney is to raise perfusion pressure — which her ventricle cannot do. Each organ's backup is the other organ. In an isolated failure, the backups are intact.

(c) One controller, two plants with opposite requirements. RAAS activation is the correct response for a kidney sensing hypoperfusion and precisely the wrong exposure for a ventricle that is stiffening from fibrosis and afterload. There is no setting of that dial that is right for both organs, which is why the therapeutic window in §33.6 is so narrow — not because either organ is badly damaged, but because they share a single volume, a single pressure, and a single hormonal axis.

(d) "Mild" is defined against an isolated norm. An ejection fraction of 48% is graded mild because, in populations of people whose kidneys, blood, lungs, and muscles are normal, that number predicts little disability. Grading systems are built one variable at a time. They therefore systematically understate risk in exactly the patients who have more than one problem — which is most patients over fifty.

The contrast case. A living kidney donor loses 50% of an organ system — severe by any anatomical measure — settles at an eGFR near 65, and has no measurable limitation in aerobic capacity or daily function. The remaining kidney hypertrophies, and crucially, nothing else in the network depends on the loss. Severity does not predict disability. Coupling does.

Question 3 — Where did the 18% rise in VO₂peak come from, with no change in ejection fraction?

Decompose it exactly. Absolute VO₂peak rose from 1.13 to 1.24 L/min (+9.7%), and body mass fell from 79.8 to 73.9 kg (−7.4%). Together: 1.097 ÷ 0.926 = 1.18, the 18% figure. So the first honest statement is that roughly 45% of the reported improvement is the denominator — she is carrying 5.9 kg less at every step. That is real and it matters functionally, and it is not a change in the cardiovascular system at all. It belongs to the digestive, endocrine, and adipose account.

The remaining 9.7% splits, by Fick, into a 2% rise in peak cardiac output and a 7.5% rise in the arteriovenous oxygen difference. Ejection fraction is unchanged because almost none of it was central. Contributing systems, at least six:

1. Muscular — the largest single contributor. Twelve weeks of aerobic training plus resistance work raises skeletal muscle capillary density by roughly 20% and mitochondrial volume density and oxidative enzyme activity by 25–40%. More capillaries means longer transit time and shorter diffusion distance; more mitochondria means a steeper intracellular oxygen gradient. Both raise extraction. This is where most of the 7.5% came from.
2. Cardiovascular, peripheral rather than central. Endothelial function improved (flow-mediated dilation 3.8% → 6.1%), so flow is better matched to demand; arterial stiffness fell (pulse wave velocity 10.8 → 9.6 m/s), so the reflected pressure wave returns later and afterload falls, improving ventricular–arterial coupling; and plasma volume expanded, raising end-diastolic volume slightly at the same filling pressure. Peak stroke volume rose modestly with no change in the ejection fraction, because both numerator and denominator rose together.
3. Nervous (autonomic). SDNN rose from 42 to 78 ms, resting heart rate fell, and one-minute heart rate recovery improved from 9 to 18 beats. Restored vagal tone and reduced resting sympathetic vasoconstrictor drive to muscle mean more of the cardiac output reaches working tissue at any given output — a redistribution gain rather than a pump gain.
4. Respiratory. Her ventilatory efficiency improved (VE/VCO₂ slope 34 → 30), so she buys the same CO₂ clearance for less respiratory muscle work — and respiratory muscles compete with locomotor muscles for cardiac output at peak effort. CPAP contributed independently: eliminating 34 apneas an hour removed a nightly sympathetic and hypoxic stimulus.
5. Blood, marrow, and gut. As the post-infarct inflammatory state resolved, hepcidin fell; her ferritin came down from 186 to 128 ng/mL while transferrin saturation rose from 11% to 21% — iron released from storage rather than iron added. Hemoglobin rose from 11.2 to 11.6 g/dL, adding about 3.6% to arterial oxygen content: a direct multiplication of every litre of output, achieved without touching the heart.
6. Endocrine, hepatic, and adipose. Improved insulin sensitivity and 4 km of daily walking recruited GLUT4 by contraction, raised fat oxidation at submaximal loads, and drove the 5.9 kg loss that supplies the denominator effect.

The general lesson, and the reason this is the last question in the book: ejection fraction is one term in a six-term product. Amara's rehabilitation improved four of the other five. A model that equates "heart failure" with "the heart" predicts that she cannot improve without a better ventricle. The integrated model predicts exactly what happened — and it is why exercise training improves symptoms and capacity in HFpEF more reliably than any drug yet tested.

Systems Integration Case File · Entry 33

Review

Level 1 · Recall

28.1 Angiotensin-converting enzyme is found in greatest quantity on the endothelium of the:

a) renal afferent arteriole    b) pulmonary capillaries    c) hepatic sinusoids    d) coronary arteries

28.2 In septic shock, the finding that most reliably distinguishes it from cardiogenic shock at the bedside is:

a) elevated lactate    b) tachycardia    c) a high mixed venous oxygen saturation    d) hypotension

28.3 Which single organ system participates in defending all nine of the regulated variables listed in Figure 33.2?

a) integumentary    b) cardiovascular    c) skeletal    d) reproductive

28.4 The acute rise in serum creatinine after starting an ACE inhibitor is caused by:

a) direct tubular toxicity    b) afferent arteriolar constriction    c) efferent arteriolar dilation    d) reduced muscle mass

28.5 Which of the following is a compensation rather than a primary abnormality in diabetic ketoacidosis?

a) anion gap of 26    b) PaCO₂ of 15 mm Hg    c) glucose of 512 mg/dL    d) bicarbonate of 6 mEq/L

28.6 Long-term arterial blood pressure is set principally by:

a) the baroreceptor reflex    b) renal pressure natriuresis    c) atrial natriuretic peptide    d) arterial compliance

28.7 In shock, the first organ to sustain measurable injury is usually the kidney. The best explanation is that:

a) the kidney receives the least blood flow at rest    b) the renal medulla already operates at a low oxygen tension and is sacrificed early by vasoconstriction    c) the kidney has no capacity for autoregulation    d) renal cells cannot use anaerobic metabolism

28.8 Contraction recruits GLUT4 to the skeletal muscle membrane by a pathway that is:

a) dependent on insulin    b) independent of insulin    c) dependent on glucagon    d) dependent on cortisol

Level 2 · Comprehension

28.9 Explain, in your own words, why "functional capacity is a product, not a sum" and give a numerical example from Amara's data.

28.10 A patient has arterial pH 7.32, PaCO₂ 30 mm Hg, HCO₃⁻ 15 mEq/L. Name the primary disorder, name the compensation, and state which organ executed each and over what timescale.

28.11 Why can a compensation be dangerous even when it is working perfectly? Give two examples from different systems.

28.12 The book claims Amara "does not have five diseases; she has one loop with several impaired nodes." Defend or refute this using at least three of her measurements.

Level 3 · Clinical Application

28.13 A 72-year-old with heart failure is admitted with breathlessness and 4 kg of weight gain. Furosemide is increased. Over three days he becomes less breathless, his weight falls 3 kg, his creatinine rises from 1.4 to 2.0 mg/dL, his potassium falls to 3.0 mEq/L and his bicarbonate rises to 34 mEq/L. Explain every one of those changes mechanistically, and say whether the treatment is working.

28.14 A 30-year-old marathon runner collapses at 32 km on a hot day. Rectal temperature 40.8 °C, confused, skin hot and dry, HR 160, BP 88/50, sodium 128 mEq/L, creatine kinase 44,000 U/L, potassium 5.9 mEq/L. Identify the failed compensations and the loops now running, and state the single most time-critical intervention.

28.15 A 55-year-old is admitted with pneumonia. On day 2: BP 86/44, HR 122, warm extremities, lactate 4.2 mmol/L, SvO₂ 79%, creatinine risen from 0.9 to 1.9 mg/dL, urine output 15 mL/h. Which term of the shock equations has failed? Why is the SvO₂ high while the lactate is also high, and what does the kidney finding tell you about the vasoconstriction hierarchy?

Level 4 · Integration and Synthesis

28.16 Construct, from first principles, the complete argument for why heart failure causes sodium retention even when total body sodium and water are already excessive. Identify every sensor, messenger, and effector, and explain why the body's error is not a malfunction.

28.17 Two patients present with a hemoglobin of 8.0 g/dL. Patient A is a 25-year-old with iron deficiency from heavy menstrual bleeding, developed over a year. Patient B is a 25-year-old who has lost the same amount of blood in the last hour from a ruptured spleen. Both have the same hemoglobin. Explain why one is walking and talking and the other is dying, using the oxygen delivery equation and at least four systems.

28.18 Choose any organ system and argue, with at least five specific mechanisms, that it would be better taught as part of a different system. Then argue the opposite. What does the exercise tell you about how physiological categories should be used?

Concept Map to Complete

Copy this onto blank paper and fill every bracket from memory, then check it. This is the whole book on one page.

                        A REGULATED VARIABLE IS OFF
                                    │
              ┌─────────────────────┴─────────────────────┐
     WHICH VARIABLE?  [ list the nine ]          WHO DEFENDS IT?
      1 [ __________ ]  4 [ __________ ]          minimum [ __ ] systems
      2 [ __________ ]  5 [ __________ ]          layered by [ ________ ]
      3 [ __________ ]  6 [ __________ ]
      7 [ __________ ]  8 [ __________ ]  9 [ __________ ]
                                    │
                    ┌───────────────┴───────────────┐
            PRIMARY FAILURE                    COMPENSATION
        moves the variable [ _____ ]      moves it [ _____ ] normal
                    │                              │
                    │                    WHAT DOES IT COST?
                    │                    [ ______________ ]
                    │                              │
                    └───────────────┬──────────────┘
                                    │
                     IS THERE A LOOP IN WHICH THE
                     COMPENSATION FEEDS THE PROBLEM?
                          │                    │
                        YES                    NO
                          │                    │
              patient [ ___________ ]   patient [ ___________ ]
              treat by [ __________ ]

     THE SIX GREAT LOOPS            THE FOUR MOVES
      a [ ____________ ]             1 [ ________ ]
      b [ ____________ ]             2 [ ________ ]
      c [ ____________ ]             3 [ ________ ]
      d [ ____________ ]             4 [ ________ ]
      e [ ____________ ]
      f [ ____________ ]     DO2 = [ __ ] × [ ____________________ ]
                             MAP = [ __ ] × [ ___ ]
                             SHOCK: four failed terms —
                               [ ______ ] [ ______ ] [ ______ ] [ ______ ]
                             ORGAN FAILURE ORDER:
                               [ __ ] → [ __ ] → [ __ ] → [ __ ] → [ __ ]

Lab / Self-Exploration

1. Run the exercise cascade on yourself. Sit quietly for five minutes and record your pulse. Now think about sprinting up a flight of stairs — do not move — and count again after thirty seconds. Then actually climb, and count immediately, at one minute, and at three minutes. You should be able to demonstrate anticipatory central command (a rise before movement), the exercise response, and vagal reactivation (the one-minute drop). Plot all five values and label each with the mechanism from Figure 33.7.
2. Demonstrate the baroreflex. Take your pulse lying down for a full minute, then stand and take it again immediately and at one and three minutes. The transient rise and partial fall is the arc in Figure 33.4 operating. Then repeat after five minutes of quiet standing in a warm room and see whether the response changes.
3. Find your own defended normals. Measure your resting heart rate and, if you have access to a cuff, your blood pressure, at the same time each morning for two weeks. Compute your mean and your own range. Compare the width of your range with the width of the population reference range. This is the n-of-1 argument in §33.10, done with your own data.
4. Build the matrix from memory. Draw the eleven-by-eleven grid of Figure 33.1 with the labels only, and fill it in without looking. Then check it. Every cell you get wrong is a connection you do not yet own — and each one is a specific, findable mechanism, not a vague association.
5. Take a case apart with the five steps. Find any clinical case report or vignette — from a course, a podcast, or a study bank — and work §33.7 on it in writing before reading the discussion. Name the variable, list the defenders, sort the findings into primary and compensatory, price each compensation, and look for the loop. Then compare your analysis with the published one and note specifically what you missed.
6. Write the one-sentence integration. Pick any two organ systems at random. In one sentence, state a mechanism by which the first changes the behavior of the second, with the direction of causation explicit. Do this ten times with ten different pairs. If you can do ten, you have the skill this book exists to teach — and you can now, honestly, close the Case File.

Key Terms

Next: the Glossary, which collects every bolded term from all thirty-three chapters into one alphabetical list, and the appendices — Appendix B for the normal laboratory reference values used throughout this book, Appendix F for the hormone table behind §33.3, and Appendix G for study strategy if pathophysiology or pharmacology is next. Amara's file stays open. You now know how to read it.