Part VI · Integration · Estimated reading time 120 minutes · Prerequisites: Chapters 2, 3, 16, 19, 22, 26
In This Chapter
- Learning Objectives
- 31.1 Body Water and the Fluid Compartments
- 31.2 Water Balance and ADH
- 31.3 Sodium: Volume, Not Concentration
- 31.4 Potassium: The Narrowest Window in the Body
- 31.5 Calcium, Phosphate, and Magnesium
- 31.6 Acid-Base: The Fundamentals
- 31.7 Respiratory and Renal Control of pH
- 31.8 The Four Primary Disorders and How to Diagnose Them
- 31.9 Advanced Topic · Integrated Disturbances
- Chapter Summary
- Case File 31 · Resolution
- Systems Integration Case File · Entry 31
- Review
- Key Terms
31. Fluid, Electrolyte, and Acid-Base Balance
The Chemistry of Homeostasis
Case File 31 — "Weak, Crampy, and Fluttering"
Amara Osei, now 45 and nine weeks after her infarction, is taking furosemide 40 mg daily for fluid overload and lisinopril 10 mg daily. Over the last week she has felt increasingly weak — "like my legs are made of wet sand" — with painful calf cramps at night and an intermittent fluttering in her chest. She has not been eating much. She has continued both medications faithfully.
Her daughter Nia, now well into her physical therapy training, recognizes the combination and brings her in.
| Laboratory | Value | Reference range |
|---|---|---|
| Sodium (Na⁺) | 133 mEq/L | 135–145 |
| Potassium (K⁺) | 2.9 mEq/L | 3.5–5.0 |
| Chloride (Cl⁻) | 92 mEq/L | 98–107 |
| Bicarbonate (HCO₃⁻) | 34 mEq/L | 22–26 |
| BUN | 32 mg/dL | 7–20 |
| Creatinine | 1.6 mg/dL | 0.6–1.1 |
| Magnesium | 1.4 mg/dL | 1.7–2.4 |
| Albumin | 3.8 g/dL | 3.5–5.0 |
| Glucose | 104 mg/dL | 70–99 |
| Arterial blood gas (room air) | Value | Reference range |
|---|---|---|
| pH | 7.49 | 7.35–7.45 |
| PaCO₂ | 47 mm Hg | 35–45 |
| HCO₃⁻ (calculated) | 34 mEq/L | 22–26 |
| PaO₂ | 88 mm Hg | 80–100 |
ECG: sinus rhythm at 92, flattened T waves, prominent U waves in the precordial leads, QU interval prolonged, occasional premature ventricular complexes.
Three questions to hold on to.
- Work out exactly what acid-base disturbance Amara has, whether it is compensated, and whether the compensation is appropriate. Show your reasoning as a sequence of steps someone else could follow.
- Furosemide acts on the thick ascending limb of the loop of Henle. But the loop is not where most potassium is finally handled — the collecting duct is. So trace exactly how a drug acting on the loop produces a potassium of 2.9 mEq/L. There is more than one mechanism, and at least one of them is not in the kidney tubule at all.
- Amara is given 80 mEq of oral potassium chloride daily for three days and her potassium is 3.0 mEq/L on repeat testing — barely moved. Why does correcting her potassium fail until her magnesium is corrected as well?
Learning Objectives
By the end of this chapter you should be able to:
- State total body water and its distribution among the fluid compartments in litres for a 70 kg adult, and explain how age, sex, and adiposity change it.
- Explain why intracellular and extracellular fluid differ in composition, tracing the difference to the Na⁺/K⁺ ATPase.
- Distinguish osmolarity, osmolality, and tonicity, calculate plasma osmolality and the osmolar gap, and explain why urea is an ineffective osmole.
- Predict, with numbers, how 1 litre of isotonic saline, of 5% dextrose in water, and of 3% saline distributes among the compartments.
- Diagram the osmoreceptor–ADH–thirst feedback loop with its thresholds, and explain why a volume stimulus overrides an osmotic one.
- State and defend the distinction that sodium determines ECF volume while water determines osmolality, and apply it to a patient with hyponatremia.
- Classify hyponatremia and hypernatremia by volume status, and explain the danger of correcting either too rapidly.
- Explain why 98% of body potassium is intracellular and why small extracellular shifts are dangerous, using the Nernst equation.
- Distinguish factors that shift potassium across membranes from those that change total body potassium.
- Explain the four determinants of potassium secretion by the collecting duct and use them to explain diuretic-induced hypokalemia.
- Explain why alkalosis lowers ionized calcium at a normal total calcium, and why magnesium depletion causes refractory hypokalemia.
- State the daily volatile and fixed acid loads and the three lines of acid-base defense with their time courses.
- State the Henderson-Hasselbalch relationship in plain language and use the Henderson equation to check an arterial blood gas for internal consistency.
- Describe how the kidney reabsorbs filtered bicarbonate, generates new bicarbonate, and excretes acid as titratable acid and ammonium.
- Apply a five-step method — pH, primary disturbance, expected compensation, anion gap, delta-delta — to interpret any arterial blood gas.
- Work through integrated disturbances including diabetic ketoacidosis, vomiting, diarrhea, decompensated COPD, salicylate overdose, and exercise-associated hyponatremia.
31.1 Body Water and the Fluid Compartments
Everything in this chapter is about a solution: what is dissolved in it, where it sits, and how tightly its properties are defended. So begin with the solvent.
How much water, and in whom
Water is 50–60% of body mass in a typical adult, but the range is wide and the reasons for the variation are entirely predictable from tissue composition:
| Tissue | Water content |
|---|---|
| Skeletal muscle | ~75% |
| Skin | ~72% |
| Blood | ~83% |
| Bone | ~22% |
| Adipose tissue | ~10% |
Adipose is essentially anhydrous. Everything else follows:
| Group | Total body water (% of mass) | Why |
|---|---|---|
| Term newborn | 75–80% | Large ECF; little fat; high surface-area-to-volume ratio |
| 1-year-old | ~65% | ECF has contracted toward adult proportions |
| Adult male | ~60% | More muscle, less fat |
| Adult female | ~50–55% | Higher proportion of adipose tissue |
| Adult with obesity | 45–50% | Adipose dilutes the water fraction |
| Adult over 75 | ~45–50% | Sarcopenia plus increased fat (Chapter 30) |
That last row is the reason Chapter 30 ends where this one begins. Adwoa's total body water is perhaps 45% of 55 kg — about 25 litres — against Nia's 55% of 58 kg, about 32 litres. A given absolute fluid loss is a larger fractional loss in an older person, before any of her renal or thirst deficits are considered.
The compartments
TOTAL BODY WATER — 70 kg adult male, ~60% of mass = 42 L
╔══════════════════════════════════════════════════════════════════════╗
║ ║
║ INTRACELLULAR FLUID (ICF) │ EXTRACELLULAR FLUID (ECF) ║
║ ~2/3 of TBW │ ~1/3 of TBW ║
║ = 28 L │ = 14 L ║
║ ┌──────────────────────────────┐ │ ┌─────────────────────────┐ ║
║ │ │ │ │ INTERSTITIAL FLUID │ ║
║ │ inside ~37 trillion cells │ │ │ ~3/4 of ECF = 10.5 L │ ║
║ │ │ │ │ (the "milieu │ ║
║ │ the largest compartment; │ │ │ intérieur" of Ch. 1) │ ║
║ │ NOT directly measurable — │ │ └─────────────────────────┘ ║
║ │ calculated as TBW − ECF │ │ ▲ ║
║ │ │ │ │ CAPILLARY WALL ║
║ └──────────────────────────────┘ │ │ (freely permeable║
║ ▲ │ │ to water, ions, ║
║ │ CELL MEMBRANE │ ▼ small solutes; ║
║ │ (freely permeable │ ┌─────────────────────────┐ ║
║ │ to WATER; ion │ │ PLASMA │ ║
║ ▼ gradients held by │ │ ~1/4 of ECF = 3.5 L │ ║
║ Na+/K+ ATPase) │ │ + 1.5 L red cells │ ║
║ │ │ = ~5 L BLOOD VOLUME │ ║
║ │ └─────────────────────────┘ ║
║ │ TRANSCELLULAR ~1–2 L ║
║ │ CSF, synovial, pleural, ║
║ │ peritoneal, aqueous, GI ║
╚══════════════════════════════════════════════════════════════════════╝
COMPOSITION — why the two sides are not the same solution (mEq/L)
ION PLASMA (ECF) INTERSTITIAL INTRACELLULAR (ICF)
────────────────────────────────────────────────────────────────
Na+ 142 145 10–14
K+ 4.2 4.4 140–150
Ca2+ (total) 5.0 2.5 <0.001 free
Mg2+ 1.8 1.5 40 (mostly bound)
────────────────────────────────────────────────────────────────
Cl- 106 117 4
HCO3- 24 27 10
PHOSPHATES 2 2 75–140
PROTEIN 16 2 40–55
────────────────────────────────────────────────────────────────
TOTAL OSMOLALITY ~285 mOsm/kg ~285 ~285 ← EQUAL
THE ONE MECHANISM BEHIND ALL OF IT:
Na+/K+ ATPase · 3 Na+ OUT / 2 K+ IN per ATP · in every cell
consumes 20–30% of resting basal metabolic rate (up to 70% in neurons)
→ Na+ is an EXTRACELLULAR ion; K+ is an INTRACELLULAR ion.
TWO SMALLER MECHANISMS:
impermeant intracellular PROTEIN and organic PHOSPHATE (fixed anions)
plasma PROTEIN (16 mEq/L) → oncotic pressure ~25 mm Hg holds plasma
volume distinct from interstitial volume across the capillary wall
Figure 31.1 — The fluid compartments as nested volumes for a 70 kg adult, with the composition contrast between intracellular and extracellular fluid.
Described: A nested diagram of total body water for a 70 kg adult male, 42 litres or about 60 percent of body mass. Two-thirds, 28 litres, is intracellular fluid inside roughly 37 trillion cells; it cannot be measured directly and is calculated as total body water minus extracellular fluid. One-third, 14 litres, is extracellular fluid, itself divided into interstitial fluid at three-quarters of the extracellular volume, or 10.5 litres, and plasma at one-quarter, or 3.5 litres; plasma plus about 1.5 litres of red cells gives a blood volume near 5 litres. A further 1 to 2 litres is transcellular fluid — cerebrospinal, synovial, pleural, peritoneal, aqueous, and gastrointestinal. The cell membrane separates intracellular from interstitial fluid and is freely permeable to water while ion gradients are held by the sodium potassium ATPase; the capillary wall separates interstitial fluid from plasma and is freely permeable to water, ions, and small solutes but not to protein. A composition table in milliequivalents per litre contrasts the compartments: sodium is 142 in plasma and only 10 to 14 inside cells, while potassium is 4.2 in plasma and 140 to 150 inside cells; chloride is 106 outside and 4 inside; bicarbonate 24 outside and 10 inside; phosphates are 2 outside and 75 to 140 inside; protein is 16 in plasma, 2 in interstitial fluid, and 40 to 55 inside cells. Total osmolality is approximately 285 milliosmoles per kilogram in every compartment — they differ in composition but not in osmolality. The single mechanism responsible is named as the sodium potassium ATPase, pumping three sodium ions out and two potassium ions in per ATP in every cell and consuming 20 to 30 percent of resting basal metabolic rate, up to 70 percent in neurons, making sodium an extracellular ion and potassium an intracellular one. Two smaller mechanisms are noted: impermeant intracellular protein and organic phosphate acting as fixed anions, and plasma protein at 16 milliequivalents per litre generating an oncotic pressure of about 25 mm Hg that keeps plasma volume distinct from interstitial volume.
Three observations from that figure will carry the whole chapter.
First: the compartments have identical osmolality and completely different composition. Water crosses every cell membrane freely through aquaporins and the lipid bilayer, so osmotic equilibrium is reached within minutes and cannot be otherwise. What differs is which solutes are on each side, and that difference is actively maintained at enormous metabolic cost.
Second: the difference traces to one pump. The Na⁺/K⁺ ATPase (Chapter 3) extrudes 3 Na⁺ and imports 2 K⁺ per ATP hydrolyzed, in essentially every cell in the body, continuously. It consumes 20–30% of resting basal metabolic rate — more than 70% in neurons. Everything downstream in this chapter — the resting membrane potential, the danger of hyperkalemia, the osmotic behavior of sodium, the ability of insulin to lower serum potassium — is a consequence of that pump.
Third: plasma and interstitial fluid are nearly identical. They differ mainly in protein (16 vs 2 mEq/L), because the capillary wall retains protein while letting water and small solutes through freely. That protein generates a colloid osmotic (oncotic) pressure of about 25 mm Hg, which is the only thing keeping plasma volume distinct from interstitial volume. When it falls — nephrotic syndrome, liver failure, severe malnutrition — the distinction fails and fluid accumulates in the interstitium as edema.
Osmolarity, osmolality, and tonicity
These three words are used interchangeably in conversation and mean different things in a clinical context.
- Osmolarity = osmoles per litre of solution (mOsm/L). Convenient for making up intravenous fluids, where you know the volume.
- Osmolality = osmoles per kilogram of solvent (mOsm/kg). What the laboratory measures and what biology actually responds to, because it is independent of temperature and of the volume occupied by dissolved solutes. In dilute body fluids the two numbers are within about 1–2% of each other. Normal plasma osmolality is 275–295 mOsm/kg.
- Tonicity (effective osmolality) = the osmolality contributed only by solutes that cannot freely cross the cell membrane, and therefore the only part that can move water.
The distinction between osmolality and tonicity is the one that matters clinically, and urea is the example that makes it clear. Urea crosses cell membranes readily, so it equilibrates on both sides and exerts no sustained osmotic force. A patient with a blood urea nitrogen of 100 mg/dL has a measured osmolality raised by about 36 mOsm/kg — and no cellular dehydration whatever, because the urea is inside the cells too. Ethanol behaves the same way. Glucose does not: in the absence of insulin it is largely excluded from muscle and fat, so hyperglycemia is effectively hypertonic and does pull water out of cells.
Calculated plasma osmolality:
Calculated osmolality (mOsm/kg) = 2 × [Na+] + glucose/18 + BUN/2.8
(mEq/L) (mg/dL) (mg/dL)
Amara: 2 × 133 + 104/18 + 32/2.8
= 266 + 5.8 + 11.4 = 283 mOsm/kg (normal 275–295)
EFFECTIVE osmolality (tonicity) = 2 × [Na+] + glucose/18
= 266 + 5.8 = 272 mOsm/kg ← mildly HYPOtonic
OSMOLAR GAP = measured − calculated. Normal < 10 mOsm/kg.
A gap > 10 means unmeasured osmoles: ethanol, methanol,
ethylene glycol, isopropanol, mannitol, or severe hyperlipidemia.
Note what those two lines say about Amara. Her measured osmolality is normal, largely because her urea is high — but urea is an ineffective osmole. Her tonicity is 272, on the low side. Tonicity, not osmolality, is what her cells experience.
Two useful factors: the ×2 on sodium accounts for its accompanying anions (mainly chloride and bicarbonate); glucose is divided by 18 and BUN by 2.8 to convert mg/dL to mmol/L using their molecular weights (180 and 28 respectively, the latter counting only the two nitrogens).
What happens when you give a litre of fluid
This is the calculation every nurse, paramedic, and physician performs mentally many times a day, and it is completely determined by the compartment model. Start with a standard 70 kg adult: TBW 42 L, ICF 28 L, ECF 14 L, plasma 3.5 L, plasma osmolality 280 mOsm/kg. Total body solute is therefore 42 L × 280 mOsm/kg = 11,760 mOsm, and it does not change unless you add solute.
Predict This
You need to raise the plasma volume of a bleeding patient by 500 mL. You have three bags: 0.9% saline, 5% dextrose in water, and 3% saline.
Before reading on, commit to an estimate: roughly how many litres of each would you need to infuse to achieve that 500 mL of plasma expansion?
(Answer: about 2 litres of 0.9% saline; about 6 litres of D5W, which is why it is never used for this; and less than 1 litre of 3% saline, which achieves it partly by taking water out of the patient's own cells. The arithmetic is below, and the reason is entirely the compartment model.)
START: TBW 42 L = ICF 28 L + ECF 14 L (plasma 3.5 L) · osm 280 mOsm/kg
total body solute = 42 × 280 = 11,760 mOsm
╔══ (A) 1 L of 0.9% NaCl "NORMAL SALINE" ═════════════════════════════════╗
║ contains Na+ 154 + Cl- 154 = 308 mOsm/L ≈ ISOTONIC with plasma ║
║ → no osmotic gradient → NO WATER MOVES ACROSS CELL MEMBRANES ║
║ ║
║ ICF 28.0 ──────────────────────────────────► 28.0 L (unchanged) ║
║ ECF 14.0 ──────────────────────────────────► 15.0 L (+1000 mL) ║
║ └── plasma 3.5 ────────────────────────► 3.75 L (+250 mL) ║
║ └── interstitial 10.5 ─────────────────► 11.25 L (+750 mL) ║
║ ║
║ ★ ONLY ~25% STAYS IN THE VESSELS. To add 500 mL of plasma volume you ║
║ must infuse ~2 L. This is the origin of the "3 : 1 rule" for ║
║ replacing blood loss with crystalloid. ║
╚════════════════════════════════════════════════════════════════════════╝
╔══ (B) 1 L of 5% DEXTROSE IN WATER (D5W) ════════════════════════════════╗
║ 278 mOsm/L in the bag, BUT the glucose is taken up and metabolized ║
║ within minutes → what you actually gave is 1 L of PURE WATER ║
║ ║
║ new TBW = 43 L · solute unchanged 11,760 · new osm = 11,760/43 ║
║ = 273.5 mOsm/kg ║
║ ICF: solute 28 × 280 = 7,840 → 7,840/273.5 = 28.67 L (+667 mL, 2/3) ║
║ ECF: solute 14 × 280 = 3,920 → 3,920/273.5 = 14.33 L (+333 mL, 1/3) ║
║ └── plasma ────────────────────────────► 3.58 L (+83 mL ONLY) ║
║ ║
║ ★ Free water distributes in proportion to compartment size, so only ║
║ ~8% reaches the plasma. USELESS as a volume expander. Its use is to ║
║ correct a pure WATER deficit (hypernatremia). ║
╚════════════════════════════════════════════════════════════════════════╝
╔══ (C) 1 L of 3% NaCl "HYPERTONIC SALINE" ═══════════════════════════════╗
║ contains Na+ 513 + Cl- 513 = 1,026 mOsm/L ≈ 3.7× plasma ║
║ ║
║ new TBW = 43 L · new solute = 11,760 + 1,026 = 12,786 ║
║ new osmolality = 12,786/43 = 297.3 mOsm/kg (↑ from 280) ║
║ ICF: solute still 7,840 → 7,840/297.3 = 26.37 L ★ ICF LOSES 1.63 L ║
║ ECF: 43 − 26.37 = 16.63 L ECF GAINS 2.63 L ║
║ └── plasma ────────────────────────────► 4.16 L (+660 mL) ║
║ ║
║ ★ A 1 L infusion expanded the ECF by 2.63 L — because 1.63 L was ║
║ RECRUITED OUT OF THE PATIENT'S OWN CELLS. That is exactly why it ║
║ works in cerebral edema, and exactly why it is dangerous. ║
╚════════════════════════════════════════════════════════════════════════╝
SUMMARY BAR (mL added to each compartment per litre infused)
ICF INTERSTITIAL PLASMA
0.9% NaCl ▪ 0 ▓▓▓▓▓▓▓ 750 ▓▓ 250
D5W ▓▓▓▓▓▓ +667 ▓▓ 250 ▪ 83
3% NaCl ◄◄◄◄ −1,630 ▓▓▓▓▓▓▓▓▓▓ 1,970 ▓▓▓▓▓▓ 660
Figure 31.2 — Where a litre of three different intravenous fluids ends up, compartment by compartment, in a 70 kg adult.
Described: Three worked calculations starting from a 70 kg adult with 42 litres of total body water — 28 intracellular and 14 extracellular, of which 3.5 is plasma — at an osmolality of 280 milliosmoles per kilogram and a total body solute of 11,760 milliosmoles. Case A, one litre of 0.9 percent sodium chloride containing 154 milliequivalents each of sodium and chloride for 308 milliosmoles per litre, is isotonic, so no water crosses cell membranes: intracellular volume is unchanged at 28 litres, extracellular volume rises by the full litre to 15 litres, of which 250 mL stays in plasma and 750 mL enters the interstitium. Only about a quarter remains intravascular, so raising plasma volume by 500 mL requires about two litres — the origin of the three-to-one rule for replacing blood loss with crystalloid. Case B, one litre of 5 percent dextrose in water, is effectively one litre of pure water because the glucose is metabolized within minutes. Total body water rises to 43 litres while solute is unchanged, so osmolality falls to 273.5. The intracellular compartment, holding 7,840 milliosmoles, expands to 28.67 litres, gaining 667 mL or two-thirds, and the extracellular compartment, holding 3,920 milliosmoles, expands to 14.33 litres, gaining 333 mL or one-third, of which only 83 mL reaches plasma — about eight percent. It is useless as a volume expander and is used instead to correct a pure water deficit. Case C, one litre of 3 percent sodium chloride containing 513 milliequivalents each of sodium and chloride for 1,026 milliosmoles per litre, adds both volume and solute: total body water becomes 43 litres and total solute 12,786, raising osmolality to 297.3. The intracellular compartment, with unchanged solute, shrinks to 26.37 litres, losing 1.63 litres, while the extracellular compartment expands by 2.63 litres to 16.63, of which 660 mL goes to plasma. A one-litre infusion therefore expands the extracellular space by more than two and a half litres because water is recruited out of the patient's own cells — the reason it works in cerebral edema and the reason it is dangerous.
Clinical Connection · Choosing an Intravenous Fluid, and Why the Choice Matters
| Fluid | Na⁺ | K⁺ | Ca²⁺ | Cl⁻ | Buffer | Osmolarity | Where it goes |
|---|---|---|---|---|---|---|---|
| 0.9% NaCl ("normal saline") | 154 | 0 | 0 | 154 | none | 308 | All ECF |
| Lactated Ringer's | 130 | 4 | 2.7 | 109 | lactate 28 | 273 | Nearly all ECF |
| Plasma-Lyte / balanced | 140 | 5 | 0 | 98 | acetate 27, gluconate 23 | 294 | All ECF |
| 5% dextrose (D5W) | 0 | 0 | 0 | 0 | none | 278 → 0 | ⅔ ICF, ⅓ ECF |
| 0.45% NaCl | 77 | 0 | 0 | 77 | none | 154 | ~⅓ ICF, ~⅔ ECF |
| 3% NaCl | 513 | 0 | 0 | 513 | none | 1,026 | ECF, plus water drawn from ICF |
| Human plasma | 140 | 4 | 5 | 100–106 | HCO₃⁻ 24 | 285–295 | — |
Three real differences, none of them cosmetic.
"Normal" saline is not normal. Its chloride concentration of 154 mEq/L is roughly 50% higher than plasma's ~104. Infusing large volumes therefore produces a hyperchloremic non-anion-gap metabolic acidosis: as chloride rises, bicarbonate must fall to preserve electroneutrality (this is the Stewart, or strong ion difference, way of seeing it, and it gives the same answer as thinking of it as a dilutional acidosis). High chloride delivery to the macula densa also causes renal afferent arteriolar vasoconstriction and reduces renal blood flow. Large randomized trials comparing balanced crystalloids with saline show small but consistent differences in kidney outcomes favoring balanced solutions in most acutely ill populations.
Lactated Ringer's contains a buffer precursor, not a buffer. Its lactate is metabolized by the liver — either oxidized or converted to glucose — and each millimole consumes a proton, generating bicarbonate. So LR is mildly alkalinizing, not acidifying, and giving it does not "raise the lactate" in a way that impairs lactate monitoring at ordinary rates. Two genuine cautions: its calcium can chelate the citrate in stored blood products (giving both through one line can cause clotting in the tubing), and its 4 mEq/L of potassium — trivial in most contexts — deserves a thought in severe hyperkalemia, though its lower chloride load usually makes it the better choice there anyway.
D5W is water. It is isotonic in the bag purely so it does not hemolyze red cells at the cannula tip. Once the glucose is metabolized, you have given free water, two-thirds of which enters cells. It is the correct fluid for a pure water deficit and the wrong fluid for hypovolemia, and giving it to a patient at risk of hyponatremia is a well-recognized route to harm.
The question to ask before hanging any bag is: what am I trying to replace — volume, free water, or both? Volume loss is replaced with an isotonic, sodium-containing fluid, because sodium is what holds fluid in the ECF (§31.3). A pure water deficit is replaced with water. Getting these backwards is one of the most common serious errors in fluid management.
Development · Why a Vomiting Infant Is an Emergency and a Vomiting Adult Usually Is Not
A term newborn is, by mass, 75–80% water — and the distribution is different too. The extracellular compartment holds roughly 40–45% of body weight at birth, against 20% in an adult, and contracts toward the adult proportion over the first year as cells grow and intracellular volume expands. Preterm infants are more extreme still, up to 85% water.
That matters because the ECF is the compartment that is lost first in vomiting, diarrhea, or fever. A newborn's larger ECF is not a reserve; it is a larger exposed surface.
Four multipliers make the situation worse.
- Surface area to volume. An infant's body surface area per kilogram is two to three times an adult's, so insensible losses through skin and lungs are proportionally far greater — roughly 30–40 mL/kg/day, and in a preterm infant under a radiant warmer, 100–200 mL/kg/day.
- Metabolic rate. Basal metabolic rate per kilogram is roughly twice an adult's, generating more metabolic waste solute that must be excreted in urine, which obliges more water.
- Immature concentrating ability. A newborn kidney can concentrate urine to only about 600–700 mOsm/kg against an adult's 1,200 — because the medullary osmotic gradient is not fully established, urea recycling is limited by the high anabolic demand for nitrogen, and collecting duct responsiveness to ADH is reduced. So the obligatory urine volume for a given solute load is roughly double.
- No autonomy. An infant cannot obtain water. Thirst, however intact, is useless without a caregiver who responds to it — a structural vulnerability shared, remarkably, with the patient in Chapter 30.
Put together: an infant turns over about 15% of total body water daily against 5–10% in an adult. A day of gastroenteritis can cost 10% of body weight. This is why pediatric dehydration is graded by weight loss, why oral rehydration solution — which exploits the sodium-glucose cotransporter to drive sodium and hence water absorption even from an inflamed gut — is one of the highest-impact interventions in global health, and why the same illness that keeps an adult home from work for two days kills a child.
Check Your Understanding 31.1
- A patient's measured serum osmolality is 320 mOsm/kg. Her sodium is 140, glucose 90, and BUN 140 mg/dL. Calculate her osmolar gap and her effective osmolality. Are her cells dehydrated?
- Why does 1 L of 0.9% saline raise plasma volume by only about 250 mL, while 1 L of 3% saline raises it by about 660 mL?
Show answers
- Calculated osmolality = 2(140) + 90/18 + 140/2.8 = 280 + 5 + 50 = 335 mOsm/kg. The measured value (320) is lower than calculated, so the osmolar gap is negative, which in practice means zero — there are no unmeasured osmoles, and the small discrepancy reflects the imprecision of the estimating equation at extreme urea values. Effective osmolality (tonicity) = 2(140) + 90/18 = 285 mOsm/kg, which is normal. Her cells are not dehydrated. Urea contributes 50 mOsm/kg to the measured number and none to the effective one, because urea crosses cell membranes freely and equilibrates on both sides. This is the single clearest demonstration of why tonicity, not osmolality, is what biology responds to — and it is why a uremic patient does not shrink her brain cells while a hyperglycemic one does.
- Because saline is isotonic and hypertonic saline is not. Isotonic saline creates no osmotic gradient across the cell membrane, so none of it enters or leaves cells; the entire litre stays in the ECF and is then divided between plasma and interstitium in the ratio the capillary wall dictates, roughly 1 : 3. Hypertonic saline adds far more solute than water, so it raises ECF tonicity, and water is then drawn out of the intracellular compartment — 1.63 L of it — until osmolality re-equilibrates. The ECF therefore expands by 2.63 L from a 1 L infusion, and the plasma share of that larger expansion is correspondingly larger. The volume is, in effect, borrowed from the patient's own cells, which is both the therapeutic principle in cerebral edema and the source of the risk.
31.2 Water Balance and ADH
The daily ledger
| Intake (typical adult, mL/day) | Output (mL/day) | ||
|---|---|---|---|
| Beverages | 1,500 | Urine | 1,500 |
| Water in food | 700 | Insensible — skin | 400 |
| Metabolic water | 300 | Insensible — respiratory tract | 300 |
| Sweat (sedentary, temperate) | 100 | ||
| Feces | 200 | ||
| Total | 2,500 | Total | 2,500 |
Metabolic water is worth a sentence because students often assume it is negligible. It is the water produced by the last step of oxidative phosphorylation, where oxygen accepts electrons and protons to form H₂O. Complete oxidation of one gram of carbohydrate yields about 0.60 g of water, of fat about 1.07 g, and of protein about 0.41 g. At ordinary intakes this totals roughly 250–350 mL/day — and it is why a hibernating or fasting animal can survive far longer without drinking than a simple accounting of intake would suggest.
Losses divide into two categories with very different clinical implications.
Obligatory losses cannot be reduced no matter how dehydrated you are:
Insensible skin + respiratory ≈ 700 mL (pure water — no solute)
Fecal water ≈ 200 mL
MINIMUM urine volume ≈ 500 mL
─────────────────────────────────────────
OBLIGATORY MINIMUM ≈ 1,400 mL/day
Why the urine has a floor:
daily solute load to excrete ≈ 600–800 mOsm (urea, Na, K, sulfate,
phosphate — largely determined by protein and salt intake)
maximum urine concentration ≈ 1,200 mOsm/kg in a healthy young adult
∴ minimum volume = 600 ÷ 1.2 = 500 mL
IN AN 80-YEAR-OLD (max urine ≈ 700 mOsm/kg, Chapter 30):
minimum volume = 600 ÷ 0.7 = 857 mL ← nearly DOUBLE the obligatory
water loss, every day
That final calculation is the quantitative statement of Adwoa's problem in Chapter 30. Her kidney cannot economize on water the way a young kidney can, so she loses more water per day just standing still — and her thirst does not tell her.
Facultative losses are the adjustable part: the urine volume above the obligatory minimum, regulated almost entirely by antidiuretic hormone.
The osmoreceptor–ADH–thirst loop
┌───────────────────────────────────────┐
│ PLASMA OSMOLALITY │
│ set point ≈ 280–285 mOsm/kg │
│ defended to within ±1–2% │
└───────────────┬───────────────────────┘
│ RISES (water loss, salt intake)
▼
╔═════════════════════════════════════════════════════════════╗
║ RECEPTOR · OSMORECEPTORS ║
║ in the OVLT and subfornical organ — CIRCUMVENTRICULAR ║
║ organs that LACK a blood-brain barrier, so they sample ║
║ plasma directly. They shrink when plasma is hypertonic; ║
║ mechanosensitive channels open; firing rate rises. ║
║ Sensitivity: detects a ~1% change in osmolality. ║
╚═══════════════╤══════════════════════════════╤══════════════╝
│ │
┌─────────────▼──────────────┐ ┌───────────▼───────────────┐
│ CONTROL CENTER · HYPOTHAL. │ │ CONTROL CENTER · CORTEX │
│ supraoptic + paraventric. │ │ conscious THIRST │
│ nuclei → axons to POSTERIOR│ │ threshold ≈ 290–295 │
│ PITUITARY │ │ mOsm/kg (HIGHER — you │
│ ADH threshold ≈ 280–285 │ │ conserve before you seek) │
└─────────────┬──────────────┘ └───────────┬───────────────┘
│ ADH (vasopressin) released │
│ ~0.4 pg/mL rise per 1 mOsm/kg│
▼ ▼
╔═══════════════════════════════════╗ ╔══════════════════════╗
║ EFFECTOR · COLLECTING DUCT ║ ║ EFFECTOR · BEHAVIOR ║
║ PRINCIPAL CELL ║ ║ drinking ║
║ V2 receptor → Gs → cAMP → PKA ║ ╚══════════╤═══════════╝
║ → AQUAPORIN-2 vesicles INSERT ║ │
║ into the apical membrane ║ │
║ → water reabsorbed down the ║ │
║ medullary osmotic gradient ║ │
║ urine 50 → 1,200 mOsm/kg ║ │
║ volume 20 L → 0.5 L per day ║ │
╚═══════════════╤═══════════════════╝ │
│ │
└──────────────┬───────────────────┘
▼
WATER RETAINED / WATER ADDED
│
▼
PLASMA OSMOLALITY FALLS ──┐
│
◄════════ NEGATIVE FEEDBACK ═════════════┘
(receptor firing falls, ADH falls, thirst stops)
╔══════════════════════════════════════════════════════════════════════╗
║ THE OVERRIDE — and the source of most clinical hyponatremia ║
║ ║
║ ARTERIAL BARORECEPTORS (carotid, aortic) and CARDIOPULMONARY ║
║ volume receptors also drive ADH. They are INSENSITIVE — a 5–10% ║
║ fall in blood volume is needed — but once triggered the response is ║
║ STEEP and OVERRIDES osmotic control entirely. ║
║ ║
║ ∴ a volume-depleted patient will retain water even though doing so ║
║ makes her HYPOtonic. The body sacrifices osmolality to defend ║
║ VOLUME — because volume failure kills in minutes and hypotonicity ║
║ takes hours. This is why heart failure, cirrhosis, vomiting, and ║
║ diuretics all cause HYPOnatremia. ║
╚══════════════════════════════════════════════════════════════════════╝
Figure 31.3 — The osmoreceptor–ADH–thirst feedback loop, with the baroreceptor override that explains most clinical hyponatremia.
Described: A negative feedback loop defending plasma osmolality at a set point of about 280 to 285 milliosmoles per kilogram, held within one to two percent. A rise in osmolality, from water loss or salt intake, is detected by osmoreceptors in the organum vasculosum of the lamina terminalis and the subfornical organ — circumventricular organs that lack a blood-brain barrier and therefore sample plasma directly. They shrink when plasma is hypertonic, opening mechanosensitive channels and raising their firing rate, and they detect a change of about one percent. Their output goes to two control centers. The first is the hypothalamic supraoptic and paraventricular nuclei, whose axons run to the posterior pituitary and release antidiuretic hormone above a threshold of 280 to 285 milliosmoles per kilogram, rising about 0.4 picograms per millilitre for each additional milliosmole. The second is the cerebral cortex, generating conscious thirst at a higher threshold of 290 to 295 — so the body conserves water before it seeks it. Antidiuretic hormone acts on collecting duct principal cells at the V2 receptor through Gs, cyclic AMP, and protein kinase A, inserting aquaporin-2 vesicles into the apical membrane so water is reabsorbed down the medullary osmotic gradient; urine concentration can range from 50 to 1,200 milliosmoles per kilogram and volume from 20 litres to half a litre per day. Thirst acts through drinking. Both effectors add or retain water, plasma osmolality falls, receptor firing declines, and the loop closes. A separate box describes the override: arterial baroreceptors in the carotid sinus and aortic arch and cardiopulmonary volume receptors also drive antidiuretic hormone release. They are insensitive, requiring a 5 to 10 percent fall in blood volume, but once triggered the response is steep and overrides osmotic control entirely, so a volume-depleted patient retains water even though this makes her hypotonic. The body sacrifices osmolality to defend volume because volume failure kills in minutes while hypotonicity takes hours — which is why heart failure, cirrhosis, vomiting, and diuretics all cause hyponatremia.
The override box deserves rereading, because it explains a pattern that otherwise looks paradoxical. Amara's sodium is 133. She is on a diuretic. She is volume-depleted, with a BUN:creatinine ratio of 20. Her body is correctly prioritizing volume over tonicity: the baroreceptor signal is driving ADH release regardless of her osmolality, so she retains free water and dilutes her sodium. Nothing is malfunctioning. A control system with two inputs and one output has resolved a conflict in favor of the more urgent variable.
Disorders of water balance, mapped onto the compartments
| Disorder | What was lost or gained | ECF | ICF | Serum Na⁺ |
|---|---|---|---|---|
| Pure water loss (fever, diabetes insipidus, no access to water) | Water only | ↓ modestly | ↓↓ (2/3 of the loss) | ↑ |
| Hypotonic fluid loss (sweat, osmotic diuresis, loop diuretic) | Water > salt | ↓↓ | ↓ | ↑ |
| Isotonic loss (hemorrhage, vomiting, drainage) | Water = salt | ↓↓↓ | unchanged | normal |
| Hypotonic hydration (water intoxication, SIADH) | Water gained | ↑ | ↑↑ | ↓ |
| Hypertonic gain (hypertonic saline, salt ingestion) | Salt > water | ↑↑ | ↓ | ↑ |
| Edema (heart failure, nephrosis, cirrhosis) | Salt and water gained, maldistributed | ↑ interstitial | unchanged | ↓ or normal |
Three of these deserve elaboration.
Dehydration. In careful usage, dehydration means a deficit of water (and therefore hypernatremia), while volume depletion means a deficit of salt and water together (and therefore ECF contraction with a variable sodium). The words are used loosely everywhere, but the distinction is exactly the one this chapter turns on and is worth preserving: one is treated with water, the other with saline.
Hypotonic hydration (water intoxication). When free water is gained faster than the kidney can excrete it — the maximum is roughly 800–1,000 mL per hour in a healthy adult, and far less when ADH is inappropriately high — plasma becomes hypotonic and water enters cells. Two-thirds of the excess ends up intracellularly. Most tissue tolerates swelling; the brain, enclosed in a rigid skull, does not. Symptoms progress from nausea and headache through confusion to seizures, brainstem herniation, and death, and they track the rate of fall far more closely than the absolute value. Causes: primary polydipsia (which requires more than about 18–20 L/day to overwhelm a normal kidney), SIADH (where the kidney cannot dilute), low-solute intake (beer potomania, "tea and toast" — with only 200 mOsm/day of solute to excrete, even a maximally dilute urine of 50 mOsm/kg can clear only 4 L), and exercise-associated hyponatremia (§31.9).
Edema. Excess fluid in the interstitial compartment. It is worth stating as a Starling problem (Chapter 19): net filtration across a capillary = K_f × [(P_c − P_i) − σ(π_c − π_i)]. Four ways to break it:
| Mechanism | Term altered | Examples |
|---|---|---|
| ↑ capillary hydrostatic pressure | P_c ↑ | Heart failure, venous obstruction, DVT, pregnancy, calcium channel blockers |
| ↓ plasma oncotic pressure | π_c ↓ | Nephrotic syndrome, liver failure, protein malnutrition, protein-losing enteropathy |
| ↑ capillary permeability | K_f ↑, σ ↓ | Inflammation, burns, sepsis, anaphylaxis |
| ↓ lymphatic drainage | removal ↓ | Lymphedema after node dissection, filariasis, tumor obstruction |
Roughly 2.5–3 litres of excess interstitial fluid must accumulate before pitting edema becomes clinically detectable — which means a patient with "mild ankle edema" is already several litres positive, and it is why daily weight is a far more sensitive monitor of fluid balance than examination. This is Amara's original indication for furosemide.
31.3 Sodium: Volume, Not Concentration
If you take one idea from this chapter, take this one.
Thread 2 · Homeostasis Is the Master Concept
Sodium content determines extracellular fluid VOLUME. Water content determines OSMOLALITY — and therefore serum sodium CONCENTRATION.
These are two separate regulated variables, controlled by two separate systems, and confusing them is the single most common conceptual error in this entire topic.
- Total body sodium is regulated by the renin-angiotensin-aldosterone system, ANP and BNP, and renal sympathetic tone, all of which are driven by sensors of volume and pressure (juxtaglomerular cells, macula densa, arterial baroreceptors, atrial stretch receptors).
- Total body water is regulated by ADH and thirst, driven by sensors of osmolality (hypothalamic osmoreceptors).
The consequence you must be able to say out loud: the serum sodium concentration tells you nothing whatsoever about total body sodium. A patient with a sodium of 125 may have a dangerously high total body sodium (decompensated heart failure, oedematous, several litres overloaded) or a dangerously low one (vomiting, diuretics, volume-depleted) or a perfectly normal one (SIADH). The number is a ratio, and you cannot read a ratio's numerator from the ratio alone.
Sodium problems are water problems. Volume problems are sodium problems. Learn that sentence backwards.
Why sodium holds the ECF together
Total body sodium in a 70 kg adult is roughly 4,000 mEq (about 58 mEq/kg): approximately half in the ECF, about 40% in bone (much of it non-exchangeable), and only about 10% inside cells. It is the dominant ECF cation at 142 mEq/L, and because the Na⁺/K⁺ ATPase continuously extrudes it, it behaves as though it were confined to the extracellular space.
Sodium plus its accompanying anions therefore constitutes roughly 90% of ECF osmotically active solute — the osmotic skeleton that holds water in the extracellular compartment. Add sodium to the body and water follows it, expanding the ECF; remove sodium and water follows it out, contracting the ECF. Water alone distributes across all compartments in proportion to their size (Figure 31.2B) and therefore expands the ECF only slightly, which is precisely why sodium and not water is the variable that determines volume.
The controls
| Signal | Stimulus | Action on sodium | Time course |
|---|---|---|---|
| Renin → angiotensin II → aldosterone | ↓ afferent arteriolar stretch; ↓ NaCl at the macula densa; β₁-sympathetic drive; (K⁺ acts directly on the adrenal) | Aldosterone ↑ ENaC and Na⁺/K⁺ ATPase in principal cells → urinary Na⁺ can fall below 10 mEq/day | Hours (genomic) |
| Angiotensin II, directly | As above | ↑ proximal tubule Na⁺/H⁺ exchange; efferent arteriolar constriction raises filtration fraction | Minutes |
| Renal sympathetic nerves | Baroreflex, exercise, hemorrhage | ↑ direct tubular Na⁺ reabsorption; renal vasoconstriction; renin release | Seconds–minutes |
| ANP (atrial) / BNP (ventricular) | Myocyte stretch from volume expansion | ↑ GFR (afferent dilation + efferent constriction); inhibits collecting duct Na⁺ reabsorption; inhibits renin, aldosterone, ADH → natriuresis | Minutes |
| ADH | Osmolality; volume depletion | Regulates water, not sodium — but by changing water it changes the sodium concentration | Minutes |
| Pressure natriuresis | ↑ renal perfusion pressure | ↑ Na⁺ excretion independent of hormones | Minutes–hours |
BNP is worth a note because Amara has had it measured: it is released by ventricular myocytes in response to wall stretch, so its concentration is a direct readout of ventricular filling pressure, which is why it works as a heart failure biomarker and why it falls when the patient is successfully decongested.
Hyponatremia
Serum Na⁺ < 135 mEq/L. It is the commonest electrolyte abnormality in hospitalized patients, and a systematic approach is essential because the causes are opposite in their treatment.
Step 1 — Is it truly hypotonic? Measure or calculate osmolality.
- Hypertonic hyponatremia: hyperglycemia or mannitol pulls water out of cells and dilutes sodium. The correction is about 2.4 mEq/L of sodium per 100 mg/dL of glucose above 100 (the older factor of 1.6 underestimates at high glucose). Worked: glucose 620, measured Na 132 → corrected Na = 132 + 2.4 × (620−100)/100 = 132 + 12.5 = 144.5. This patient is hypernatremic, not hyponatremic, and has a large free water deficit hiding behind a reassuring-looking number.
- Isotonic ("pseudo") hyponatremia: severe hyperlipidemia or paraproteinemia displaces plasma water, so a flame-photometry sodium reads low while the sodium concentration in plasma water is normal. A laboratory artifact with direct ion-selective electrodes largely abolished — but still worth knowing.
Step 2 — Assess volume status, because that determines both the cause and the treatment.
| Volume status | Mechanism | Causes | Urine Na⁺ | Treatment |
|---|---|---|---|---|
| Hypovolemic | Salt and water lost; ADH driven by volume; water replaced without salt | Thiazides (the classic), vomiting, diarrhea, third-spacing, adrenal insufficiency, cerebral salt wasting | > 20 if renal loss; < 20 if extrarenal | Isotonic saline |
| Euvolemic | Water retained without salt | SIADH, hypothyroidism, glucocorticoid deficiency, primary polydipsia, low solute intake | Usually > 20; urine osm > 100 in SIADH | Fluid restriction; treat cause |
| Hypervolemic | Total body sodium is high, but total body water is higher; low effective arterial blood volume drives ADH | Heart failure, cirrhosis, nephrotic syndrome, advanced CKD | < 20 (except CKD) | Salt and fluid restriction; diuretics |
Note the trap in the third row: these patients are overloaded with sodium and hyponatremic at the same time. Giving them saline "for the low sodium" makes them worse. This is the clinical payoff of the volume-versus-concentration distinction.
Step 3 — Correct at a safe rate. Symptoms come from cerebral edema, and correction risks the opposite injury.
Clinical Connection · Why Correcting Sodium Too Fast Injures the Brain — in Both Directions
The brain adapts to a slowly changing tonicity, and the adaptation is what makes rapid correction dangerous.
In chronic hyponatremia (present more than ~48 hours), brain cells defend their volume by extruding organic osmolytes — myo-inositol, taurine, glutamate, glutamine, creatine — over 24 to 48 hours. The brain is therefore no longer swollen at a sodium of 118; it has re-equilibrated at a lower intracellular solute content. If serum sodium is then raised quickly, water leaves brain cells faster than osmolytes can be resynthesized (which takes days), the cells shrink, and osmotic demyelination syndrome follows — classically central pontine myelinolysis, with oligodendrocyte injury and demyelination in the basis pontis and often extrapontine sites. The clinical picture is devastating and characteristically delayed by two to seven days after an apparently successful correction: dysarthria, dysphagia, quadriparesis, and in severe cases a locked-in state.
Correction limit: no more than 8 mEq/L in 24 hours (some guidelines allow 10–12; high-risk patients — alcohol use disorder, malnutrition, hypokalemia, liver disease, sodium < 105 — warrant 4–6). Note the asymmetry: it is safe to raise sodium by 4–6 mEq/L rapidly in the first hours if the patient is seizing, because that is enough to relieve cerebral edema. The limit is on the 24-hour total, not on the initial rate.
In chronic hypernatremia the brain does the reverse: over hours to days it generates idiogenic osmoles to hold its volume against a hypertonic plasma. Lowering serum sodium quickly then leaves the brain relatively hypertonic to the plasma, water rushes in, and cerebral edema and seizures follow. Correction limit: no more than 10–12 mEq/L in 24 hours, roughly 0.5 mEq/L per hour.
The tool for planning either correction is the Adrogué-Madias estimate of the effect of one litre of a chosen fluid:
Δ[Na+] per litre infused = (infusate Na+ + infusate K+) − serum Na+
───────────────────────────────────────
total body water + 1
Worked, for Adwoa from Chapter 30: 55 kg, elderly female, so TBW ≈ 0.45 × 55 = 24.75 L. Serum Na 148. If you infuse D5W (Na 0): Δ[Na⁺] per litre = (0 − 148)/(24.75 + 1) = −5.7 mEq/L per litre. To lower her sodium from 148 to 140 — a fall of 8 mEq/L, safely inside the 10–12 limit — requires about 1.4 L, given over 24 hours, plus replacement of her ongoing obligatory losses.
Cross-check with the free water deficit formula: deficit = TBW × ([Na⁺]/140 − 1) = 24.75 × (148/140 − 1) = 24.75 × 0.0571 = 1.41 L. The two methods agree, which is the point of doing both.
The unifying principle: in both hyponatremia and hypernatremia, the brain injury is caused by the correction, not by the disorder. Slow is safe; fast is not; and the reason is that cellular osmolyte adjustment takes days while plasma sodium can be changed in hours.
Hypernatremia
Serum Na⁺ > 145 mEq/L. It always means a deficit of water relative to sodium, and — because thirst is such a powerful defense — it almost always requires that thirst be impaired or that water be unobtainable. That is why hypernatremia is a disease of infants, of intubated or sedated patients, of people with dementia, and of the frail elderly.
| Category | Mechanism | Examples |
|---|---|---|
| Pure water loss | Water without solute | Insensible losses with fever; central or nephrogenic diabetes insipidus |
| Hypotonic fluid loss | Water lost in excess of sodium | Osmotic diuresis (hyperglycemia, mannitol, high-protein feeds), loop diuretics, sweating, vomiting, diarrhea in children |
| Sodium gain (rare) | Solute added | Hypertonic saline, sodium bicarbonate in resuscitation, salt ingestion, hypertonic feeds |
Symptoms are neurological — lethargy, irritability, weakness, and in severe or rapid cases seizures and intracranial hemorrhage from shrinking brain tissue tearing bridging veins. In older adults the presentation is often simply confusion (Chapter 30), which is exactly what happened to Adwoa.
Aging · Why Every Disorder in This Chapter Falls Hardest on the Oldest Patients
Chapter 30 established the individual deficits. Assemble them here, and the answer to "why is the elderly patient always the one with the electrolyte problem?" becomes arithmetic rather than assertion.
| Age-related change | Chapter 30 § | Consequence in this chapter |
|---|---|---|
| Total body water 45–50% rather than 60% | §30.6 | A given absolute loss is a larger fractional loss; drug volumes of distribution shift |
| Blunted thirst | §30.7 | The deficit is not self-corrected — the loop's behavioral effector is offline |
| Max urine osmolality 700–800, not 1,200 | §30.7 | Obligatory urine volume nearly doubles: 857 mL/day rather than 500 |
| Impaired diluting ability, ↓ GFR | §30.7 | Free water cannot be excreted quickly either — hyponatremia risk rises symmetrically |
| Renin and aldosterone ↓ 30–50% | §30.6 | Slower renal sodium conservation; salt wasting under restriction; more hyperkalemia with ACE inhibitors |
| GFR ↓ with a NORMAL creatinine | §30.7 | Diuretics, ACE inhibitors, NSAIDs, and every renally cleared drug accumulate |
| Blunted β-adrenergic response; stiff ventricle | §30.5 | Volume depletion is poorly tolerated (preload-dependent) and volume overload is poorly tolerated (stiff ventricle). A narrow ledge on both sides |
| Polypharmacy | §30.8 | Thiazides, SSRIs, PPIs, NSAIDs, and laxatives are all direct causes of the disorders in this chapter |
Adwoa, day 2, worked as a fluid problem. Sodium 148, BUN 38, creatinine 1.4 from a baseline 0.9, BUN:creatinine ratio 27 — a hypertonic volume deficit with prerenal azotemia. She was on a thiazide (ongoing sodium and water loss), took in about 200 mL over eighteen hours (intake failure), could not concentrate her urine efficiently (obligatory loss elevated), and was not thirsty (correction failure). Her free water deficit, computed in the sidebar above, is about 1.4 L — a quantity a healthy 24-year-old would replace without noticing, over a single afternoon, in response to a sensation she would not even bother to name.
The rest of her presentation follows: the hypertonicity and reduced cerebral perfusion pushed a brain with minimal reserve past its threshold, and her diphenhydramine — cleared by a liver and kidney working at reduced capacity, acting on a brain with increased sensitivity — did the rest. Every element of her delirium is on the table above.
Check Your Understanding 31.3
- Two patients both have a serum sodium of 126 mEq/L. Patient A has jugular venous distension, pitting edema to the knees, and crackles. Patient B has dry mucous membranes, flat neck veins, and orthostatic hypotension. Which has more total body sodium, and how should each be treated?
- A patient with a chronic sodium of 112 mEq/L is corrected to 132 over 18 hours and appears to improve. Four days later she develops dysarthria, dysphagia, and quadriparesis. What happened, and what should have been done?
Show answers
- Patient A has far more total body sodium — probably several hundred milliequivalents in excess — despite the identical serum concentration. She has hypervolemic hyponatremia from heart failure: her total body sodium is high, her total body water is higher, and the low sodium concentration reflects that ratio. Her low effective arterial blood volume (a failing heart delivers inadequate flow despite a full circulation) triggers the baroreceptor ADH override in Figure 31.3, so she retains free water. Treatment is sodium and fluid restriction plus a diuretic — remove both, with more water than sodium. Giving her saline would worsen both her congestion and, eventually, her sodium. Patient B has hypovolemic hyponatremia: total body sodium is low. Treatment is isotonic saline, which restores volume, switches off the non-osmotic ADH drive, and allows the kidney to excrete the retained free water — often correcting the sodium faster than intended, which is precisely why these patients need close monitoring.
- Osmotic demyelination syndrome from over-rapid correction. Her sodium rose 20 mEq/L in 18 hours, more than double the 8 mEq/L per 24 hours limit. Because her hyponatremia was chronic, her brain cells had extruded organic osmolytes to normalize their volume; raising serum sodium quickly then drew water out of cells faster than osmolytes could be resynthesized, injuring oligodendrocytes and producing demyelination, classically in the pons. The delayed onset of two to seven days after an apparently good correction is characteristic and is why the initial improvement was falsely reassuring. What should have been done: a target of no more than 8 mEq/L in 24 hours (and, given a starting sodium of 112, arguably 4–6), with sodium measured every 2–4 hours during active correction; a small bolus of hypertonic saline only if she had been seizing, aiming for a 4–6 mEq/L rise to relieve cerebral edema; anticipation of auto-correction — if the cause was volume depletion or a stopped thiazide, the kidney will dump free water on its own once ADH switches off, and the sodium will rise fast without any further treatment; and, if over-correction occurred, deliberate re-lowering with free water and desmopressin, which is an established rescue strategy.
31.4 Potassium: The Narrowest Window in the Body
Why 2% of the body's potassium causes all the trouble
Total body potassium in a 70 kg adult is about 3,500 mEq (roughly 50 mEq/kg). Of that, 98% is intracellular — about 3,430 mEq — and 2%, roughly 70 mEq, is extracellular. Of the extracellular fraction, the plasma holds only about 15–20 mEq in total.
Sit with those numbers, because the whole of potassium physiology follows from them. The extracellular pool is so small that a shift of just 1% of intracellular potassium into the ECF — about 34 mEq — would roughly double the serum concentration. Conversely, a serum potassium of 2.9 mEq/L, as Amara has, can accompany a total body deficit of 200–400 mEq or more, because the ECF is the visible tip of a very large iceberg and the concentration is a poor guide to the total.
Two consequences:
- Serum potassium is exquisitely sensitive to shifts and insensitive to total body content. A patient can have a normal serum potassium and be profoundly depleted (diabetic ketoacidosis, §31.9), or a raised serum potassium with normal total body content (acidosis, cell lysis).
- Small changes in serum potassium have large electrical consequences, because the ratio of intracellular to extracellular potassium sets the resting membrane potential.
The membrane potential, revisited
From Chapter 11: the resting membrane potential of an excitable cell is dominated by potassium, because the resting membrane is far more permeable to K⁺ than to any other ion. The Nernst equation gives the equilibrium potential at 37 °C:
E_K = 61 × log ( [K+]outside / [K+]inside ) mV
Work it three times, holding intracellular K⁺ at 140 mEq/L:
| [K⁺]outside | E_K | Effect on resting membrane potential |
|---|---|---|
| 2.5 mEq/L | 61 × log(2.5/140) = −107 mV | Hyperpolarized — further from threshold |
| 4.0 mEq/L (normal) | 61 × log(4/140) = −94 mV | Normal (measured resting potential ≈ −90 mV) |
| 6.0 mEq/L | 61 × log(6/140) = −83 mV | Depolarized by ~11 mV |
| 8.0 mEq/L | 61 × log(8/140) = −76 mV | Depolarized by ~18 mV |
Hyperkalemia depolarizes. The first effect of mild depolarization is increased excitability — the cell sits closer to threshold. But sustained depolarization causes voltage-gated sodium channels to enter their inactivated state, from which they cannot open until the membrane repolarizes. As more channels inactivate, the upstroke of the action potential becomes smaller and slower, conduction slows, and eventually the tissue becomes inexcitable. That biphasic sequence — briefly more excitable, then progressively less — is exactly what the ECG shows as potassium climbs.
Hypokalemia hyperpolarizes — but the cardiac consequence is subtler and, at first sight, backwards. A hyperpolarized cell is harder to excite, which explains the skeletal muscle weakness. In cardiac muscle, however, the conductance of the inward-rectifier potassium current I_K1 is proportional to the square root of extracellular potassium. So when extracellular potassium falls, cardiac potassium conductance falls, repolarization slows, the action potential lengthens, and the QT/QU interval prolongs. That creates the substrate for early afterdepolarizations and torsades de pointes. Hypokalemia is therefore paradoxically arrhythmogenic despite hyperpolarizing the cell, and this is the mechanism behind the U waves on Amara's ECG.
What moves potassium, and what changes how much there is
TOTAL BODY K+ ≈ 3,500 mEq · diet 50–100 mEq/day · 90% renal loss
╔═══════════════════════════════════════════════════════════════════════╗
║ INTRACELLULAR EXTRACELLULAR ║
║ ~3,430 mEq (98%) ~70 mEq (2%) ║
║ [K+] ≈ 140 mEq/L [K+] ≈ 4 mEq/L ║
║ ┌───────────────────────────┐ ┌──────────────────┐ ║
║ │ muscle holds ~75% │ Na+/K+ ATPase │ plasma alone │ ║
║ │ liver, RBC, bone rest │ ◄══ 2 K+ IN ════│ holds only │ ║
║ │ │ ═══ 3 Na+ OUT ═►│ ~15–20 mEq ! │ ║
║ └───────────────────────────┘ └──────────────────┘ ║
╚═══════════════════════════════════════════════════════════════════════╝
▲ RATIO 35 : 1 sets E_K = −94 mV = the resting membrane potential ▲
┌─────────────────────────────────────────────────────────────────────────┐
│ (A) SHIFTS — change serum K+ WITHOUT changing total body K+ │
├──────────────────────── K+ INTO CELLS ─┬─ K+ OUT OF CELLS ──────────────┤
│ INSULIN (↑ Na+/K+ ATPase activity) │ INSULIN DEFICIENCY │
│ β2-AGONISTS (epinephrine, albuterol; │ α-ADRENERGIC stimulation │
│ cAMP → ↑ pump) │ β-BLOCKADE │
│ ALKALOSIS (H+ leaves cells to buffer; │ ACIDOSIS — mineral (HCl, NH4Cl)│
│ K+ enters for electroneutrality) │ ~0.3–0.6 mEq/L per 0.1 pH │
│ ↑ CELL BUILDING (refeeding, B12 │ ORGANIC acidoses (lactic, │
│ therapy, GM-CSF) │ keto) shift K+ much LESS, │
│ HYPOTHERMIA │ because the anion crosses │
│ │ the membrane with the H+ │
│ │ HYPERTONICITY (hyperglycemia, │
│ │ mannitol): water leaves the │
│ │ cell, [K+]in rises, K+ exits │
│ │ ~0.4–0.8 mEq/L per 100 mg/dL │
│ │ CELL LYSIS (rhabdomyolysis, │
│ │ tumor lysis, hemolysis) │
│ │ EXERCISE (K+ from contracting │
│ │ muscle; up to 6–8 mEq/L at │
│ │ maximal effort; normalizes │
│ │ within minutes of stopping) │
│ │ DIGOXIN TOXICITY (pump blocked)│
└────────────────────────────────────────┴────────────────────────────────┘
┌─────────────────────────────────────────────────────────────────────────┐
│ (B) EXTERNAL BALANCE — change TOTAL BODY K+ │
│ │
│ IN: diet 50–100 mEq/day │
│ OUT: kidney ~90% · gut ~10% (colon secretes K+; ↑ in CKD) │
│ │
│ RENAL HANDLING │
│ filtered ~800 mEq/day │
│ PROXIMAL TUBULE reabsorbs ~65% (paracellular + solvent drag) │
│ THICK ASC. LIMB reabsorbs ~25% (NKCC2; most is RECYCLED back │
│ to the lumen through ROMK to │
│ sustain the lumen-POSITIVE │
│ potential) │
│ ★ COLLECTING DUCT SECRETES — and this is where regulation lives │
└─────────────────────────────────────────────────────────────────────────┘
┌─────────────────────────────────────────────────────────────────────────┐
│ (C) THE FOUR LEVERS ON COLLECTING-DUCT K+ SECRETION (principal cell) │
│ │
│ LUMEN PRINCIPAL CELL BLOOD │
│ │ │
│ │◄── Na+ ── ENaC ──┐ ┌── Na+/K+ ATPase ──► │
│ │ (makes lumen │ │ 3 Na+ out │
│ │ NEGATIVE) ▼ │ 2 K+ in ◄── K+ │
│ │◄── K+ ─── ROMK ──┤ [K+] high inside ────┘ │
│ │◄── K+ ─── BK ────┘ because the pump keeps loading it │
│ │ (flow-activated) │
│ │
│ 1 · ALDOSTERONE ↑ ENaC, ↑ Na+/K+ ATPase, ↑ ROMK │
│ 2 · DISTAL Na+ DELIVERY more Na+ through ENaC → lumen more NEGATIVE │
│ → stronger electrical pull on K+ │
│ 3 · TUBULAR FLOW RATE sweeps K+ downstream (keeps the chemical │
│ gradient steep) AND bends the cilium → │
│ Ca2+ → opens BK channels │
│ 4 · NON-REABSORBABLE HCO3-, ketoanions, penicillin, hippurate │
│ LUMINAL ANION → lumen more negative → more K+ secreted │
│ │
│ (α-INTERCALATED CELLS reabsorb K+ via H+/K+-ATPase — upregulated in │
│ K+ depletion, and the reason K+ depletion also causes ALKALOSIS) │
└─────────────────────────────────────────────────────────────────────────┘
Figure 31.4 — Potassium distribution, the shifts that move it between compartments, and the four levers controlling its renal secretion.
Described: A four-part diagram. The first panel shows total body potassium of about 3,500 milliequivalents divided between an intracellular pool of about 3,430 milliequivalents, 98 percent, at a concentration near 140 milliequivalents per litre with muscle holding about three quarters of it, and an extracellular pool of about 70 milliequivalents, 2 percent, at 4 milliequivalents per litre, of which plasma alone holds only 15 to 20 milliequivalents. The sodium potassium ATPase maintains the split, pumping two potassium in and three sodium out, and the resulting 35 to 1 ratio sets the potassium equilibrium potential at minus 94 millivolts. The second panel lists shifts that change serum potassium without changing total body potassium. Moving potassium into cells: insulin, beta-2 agonists such as epinephrine and albuterol acting through cyclic AMP on the pump, alkalosis, states of rapid cell building such as refeeding, and hypothermia. Moving potassium out of cells: insulin deficiency, alpha adrenergic stimulation, beta blockade, mineral acidosis at roughly 0.3 to 0.6 milliequivalents per litre per 0.1 pH unit — with the note that organic acidoses such as lactic acidosis and ketoacidosis shift potassium much less because the anion crosses the membrane with the proton — hypertonicity from hyperglycemia or mannitol at 0.4 to 0.8 milliequivalents per 100 mg/dL of glucose, cell lysis from rhabdomyolysis, tumor lysis or hemolysis, exercise which can raise plasma potassium to 6 to 8 milliequivalents per litre at maximal effort and normalizes within minutes, and digoxin toxicity which blocks the pump. The third panel covers external balance: intake of 50 to 100 milliequivalents daily, with about 90 percent excreted renally and 10 percent through the gut. Of roughly 800 milliequivalents filtered daily, the proximal tubule reabsorbs about 65 percent and the thick ascending limb about 25 percent through the sodium-potassium-two-chloride cotransporter, though most of that is recycled back into the lumen through ROMK channels to sustain the lumen-positive potential; the collecting duct secretes potassium and is where regulation occurs. The fourth panel diagrams a collecting duct principal cell: the basolateral sodium potassium ATPase loads the cell with potassium, apical epithelial sodium channels admit sodium and make the lumen negative, and potassium exits apically through ROMK and through flow-activated BK channels. Four levers control secretion: aldosterone, which increases the sodium channel, the pump, and ROMK; distal sodium delivery, since more sodium through the channel makes the lumen more negative and pulls potassium out more strongly; tubular flow rate, which sweeps potassium downstream keeping the chemical gradient steep and bends the primary cilium to raise calcium and open BK channels; and non-reabsorbable luminal anions such as bicarbonate, ketoanions, penicillin, and hippurate, which make the lumen more negative. A closing note records that alpha intercalated cells reabsorb potassium through a hydrogen-potassium ATPase that is upregulated in potassium depletion, which is why potassium depletion also causes alkalosis.
Resolving Case File question 2 — how a loop diuretic empties the potassium stores
Furosemide inhibits the Na⁺/K⁺/2Cl⁻ cotransporter (NKCC2) in the apical membrane of the thick ascending limb. That is the whole of its direct action. The question asks how a drug acting there produces a potassium of 2.9 — and the answer is that almost all of the potassium loss happens downstream of the drug's site of action, by five mechanisms operating together.
Mechanism 1 — the direct loss is real but small, and partly self-cancelling. The thick ascending limb does reabsorb about 25% of the filtered potassium load through NKCC2, so blocking the transporter does cause some direct loss. But most of the potassium that NKCC2 brings into the cell is immediately recycled back into the lumen through apical ROMK channels, because that recycling is what generates the lumen-positive transepithelial potential the segment depends on. Blocking NKCC2 therefore removes both the uptake and much of the recycling. Net direct loss: modest.
Mechanism 2 — increased distal sodium delivery. This is the dominant mechanism. Sodium not reabsorbed in the thick ascending limb flows on to the distal convoluted tubule and collecting duct. There, principal cells reabsorb it through ENaC. Sodium entering through ENaC is a cation leaving the lumen, so the lumen becomes more electrically negative — and the lumen's negativity is the electrical force pulling potassium out of the cell through ROMK. More sodium delivered means more sodium through ENaC means a more negative lumen means more potassium secreted. Lever 2 in Figure 31.4C.
Mechanism 3 — increased tubular flow. The same blocked reabsorption that delivers extra sodium also delivers extra water. High flow does two things at once: it sweeps secreted potassium downstream so the luminal concentration stays low and the chemical gradient for further secretion stays steep, and it bends the principal cell's primary cilium, raising intracellular calcium and opening BK (maxi-K) channels that are essentially silent at resting flow rates. Lever 3. Mechanisms 2 and 3 are multiplicative, not additive, which is why loop diuretics waste potassium so effectively.
Mechanism 4 — RAAS activation, and this one is not in the kidney tubule at all. Furosemide depletes ECF volume. Volume depletion is detected by afferent arteriolar stretch receptors, by the macula densa (which now senses less NaCl, because NKCC2 in the macula densa is the same transporter the drug blocks — so the signal is doubly amplified), and by arterial baroreceptors driving renal sympathetic outflow. All three release renin. Renin → angiotensin II → aldosterone, which inserts more ENaC, more Na⁺/K⁺ ATPase, and more ROMK into principal cells. Lever 1, superimposed on levers 2 and 3.
This is where Amara's second drug matters. Her lisinopril blocks angiotensin-converting enzyme and should blunt aldosterone release. It does — partially. But the volume-depletion stimulus is powerful, aldosterone "breakthrough" during chronic ACE inhibition is well described, and potassium itself directly stimulates the adrenal zona glomerulosa only when it is high, not when it is low. Her ACE inhibitor is pushing her potassium up while her furosemide pushes it down, and the furosemide is winning by a wide margin. Note also that lisinopril is contributing to her rising creatinine: by dilating the efferent arteriole in a volume-depleted patient it lowers glomerular filtration pressure — a hemodynamic effect, not tubular injury.
Mechanism 5 — the alkalosis feeds back on the potassium, and vice versa. The metabolic alkalosis (§31.8) does two things. It shifts potassium into cells, lowering the measured serum value further. And it delivers more bicarbonate to the distal nephron, where bicarbonate acts as a poorly reabsorbable luminal anion — lever 4 — increasing lumen negativity and driving still more potassium secretion. Meanwhile the potassium depletion itself perpetuates the alkalosis, by shifting H⁺ into cells, by stimulating the H⁺/K⁺-ATPase of α-intercalated cells (which reclaims potassium at the price of secreting acid), and by increasing renal ammoniagenesis. Hypokalemia and metabolic alkalosis are mutually reinforcing, which is why treating one without the other fails.
And mechanism 6, which sets up question 3. The lumen-positive potential of the thick ascending limb is also what drives the paracellular reabsorption of calcium and magnesium through the claudin-16/19 pore. Furosemide abolishes that potential, so magnesium and calcium are wasted along with everything else. Amara's magnesium of 1.4 mg/dL is a direct consequence of her furosemide — and, as §31.5 shows, it is the reason her potassium will not come up.
Hypokalemia and hyperkalemia
| Hypokalemia (< 3.5 mEq/L) | Hyperkalemia (> 5.0 mEq/L) | |
|---|---|---|
| Shift causes | Insulin, β₂-agonists, alkalosis, refeeding, hypothermia, hypokalemic periodic paralysis | Acidosis (mineral), insulin deficiency, β-blockade, hypertonicity, digoxin toxicity, succinylcholine, cell lysis |
| Loss / gain causes | Diuretics (loop and thiazide), vomiting via renal loss, diarrhea, hyperaldosteronism, magnesium depletion, RTA types 1 and 2, amphotericin, Bartter and Gitelman syndromes, poor intake | Reduced excretion: AKI/CKD, ACE inhibitors, ARBs, spironolactone, trimethoprim, heparin, NSAIDs, type 4 RTA, hypoaldosteronism; excess intake with impaired excretion |
| First check | Magnesium, and whether the patient is alkalotic | Pseudohyperkalemia: tourniquet, fist clenching, hemolysed sample, extreme leukocytosis or thrombocytosis |
| Muscle | Proximal weakness → respiratory failure; cramps; ileus; rhabdomyolysis when severe | Weakness, paresthesia, ascending paralysis |
| Kidney | Nephrogenic diabetes insipidus (AQP2 downregulated) → polyuria and polydipsia | — |
| ECG | T-wave flattening, ST depression, U waves, long QU, PVCs, torsades | Peaked T waves → PR prolongation → P-wave loss → wide QRS → sine wave |
| Treatment logic | Replace K⁺ and Mg²⁺; ≤ 10 mEq/h peripherally; oral where possible; treat the cause; a fall of 1 mEq/L below 4.0 implies a deficit of roughly 200–400 mEq | Stabilize → shift → remove (see below) |
A note on vomiting and potassium, because it is almost always taught wrongly. Gastric fluid contains only about 5–10 mEq/L of potassium, so the potassium lost in the vomit itself is trivial. The hypokalemia of vomiting is overwhelmingly renal: the alkalosis delivers bicarbonate to the distal nephron as a non-reabsorbable anion, and the volume depletion drives aldosterone. The kidney does the losing; the stomach only starts it.
Clinical Connection · Hyperkalemia: Three Steps, in This Order
Severe hyperkalemia — a serum potassium above about 6.5 mEq/L, or any level with ECG changes — is one of the few true minute-to-minute emergencies in medicine, and the treatment sequence is worth understanding mechanistically rather than memorizing.
1 · STABILIZE the membrane — IV calcium. 10 mL of 10% calcium gluconate (or calcium chloride through a central line), over 2–3 minutes, working within 1–3 minutes and lasting 30–60 minutes. It does not lower the potassium at all. Hyperkalemia raises the resting membrane potential toward threshold, narrowing the gap between them; calcium raises the threshold potential, restoring the gap and re-establishing normal excitability. It is a purely electrical intervention that buys time. (The old teaching that calcium is contraindicated in digoxin toxicity is now considered overstated, but caution persists.)
2 · SHIFT potassium into cells — temporary, and it treats the number rather than the problem. - Insulin 10 units IV with 25 g of dextrose — the most reliable agent. Onset 10–20 minutes, peak at 30–60, duration 4–6 hours; lowers K⁺ by 0.5–1.2 mEq/L. Mechanism: insulin directly increases Na⁺/K⁺ ATPase activity and abundance. Hypoglycemia is common afterwards — the glucose is cleared faster than the insulin — so monitor and consider an infusion. - Nebulized albuterol 10–20 mg (four to eight times an asthma dose): β₂ → Gs → cAMP → pump. Onset ~30 minutes, lowers K⁺ by 0.5–1.0 mEq/L. Additive with insulin. About a quarter of patients do not respond. - Sodium bicarbonate: useful only if there is a genuine metabolic acidosis; unreliable as a standalone potassium-lowering agent.
3 · REMOVE potassium from the body — the only step that actually treats it. - Loop diuretic with saline, if the patient makes urine (mechanisms 2 and 3 above, now used deliberately). - Potassium binders — patiromer, sodium zirconium cyclosilicate — which exchange potassium in the gut lumen. Onset hours; useful for ongoing control. - Hemodialysis — definitive, and the answer in oliguric renal failure.
Note the structure of the logic: one intervention protects the tissue, one hides the number, and one fixes the problem, and you do them in that order because they act on different time scales. Also note that Amara is at risk of hyperkalemia in the future, not now: as her chronic kidney disease advances, her lisinopril's effect on aldosterone will stop being a helpful counterweight to the furosemide and start being a hazard on its own.
Imaging · The ECG Is the "Imaging" of Potassium
There is little to image in this chapter — but there is one investigation that renders the intracellular electrical consequences of an extracellular ion concentration in real time, at the bedside, in seconds. It is the reason an ECG is obtained before the potassium result returns in any patient at risk.
[K+] ECG MORPHOLOGY (schematic, lead V3) MECHANISM
mEq/L
────────────────────────────────────────────────────────────────────────────
8.0 ╭──╮ ╭──╮ ╭──╮ Na+ channels largely
SINE │ ╰──────╯ ╰──────╯ ╰── INACTIVATED; QRS merges
WAVE sinusoidal — no distinct P, QRS or T with T. Pre-arrest.
────────────────────────────────────────────────────────────────────────────
7.5 ▁▁▁╱▔▔▔╲▁▁▁ no P wave Atrial myocardium is
WIDE wide slurred QRS, tall peaked T inexcitable before
QRS > 120 ms ventricular; conduction
slows everywhere
────────────────────────────────────────────────────────────────────────────
6.5 ‿ ▁╱▏╲▁ ╱▔╲ PR lengthens; P flattens.
PR ↑ small P · long PR · peaked T Depolarization threshold
harder to reach as Na+
channels inactivate
────────────────────────────────────────────────────────────────────────────
5.5 ╱▔╲ ▁╱▏╲▁ ╱▔╲ TALL, NARROW, PEAKED,
PEAKED normal P · normal QRS · TENTED T symmetrical T with a
T T narrow-based and "tented" narrow base — repolarization
is FASTER because I_K1
conductance ∝ √[K+]o
────────────────────────────────────────────────────────────────────────────
4.0 ╱▔╲ ▁╱▏╲▁ ╭─╮ NORMAL
NORMAL P · QRS · rounded T · no U E_K = −94 mV
────────────────────────────────────────────────────────────────────────────
3.0 ╱▔╲ ▁╱▏╲▁ ╭╮ ╭╮ T FLATTENS; U wave
U WAVE P · QRS · flat T · U wave appears appears after the T.
ST segment depresses I_K1 conductance FALLS, so
repolarization is SLOWED
────────────────────────────────────────────────────────────────────────────
2.5 ╱▔╲ ▁╱▏╲▁ ‿ ╭──╮ U EXCEEDS T. Long QU.
U > T T inverted or absent · prominent U Substrate for early
marked ST depression afterdepolarizations →
TORSADES DE POINTES
────────────────────────────────────────────────────────────────────────────
2.0 ╱▔╲ ▁╱▏╲▁ ╭────╮ ╭╮╭╮╭╮ Sustained polymorphic VT
TdP fused T-U · polymorphic VT Digoxin toxicity greatly
potentiated at any low K+
────────────────────────────────────────────────────────────────────────────
★ AMARA sits at 2.9: flat T, prominent U, long QU, occasional PVCs.
★ CAUTION: the ECG is SPECIFIC but NOT SENSITIVE. A normal ECG does not
exclude dangerous hyperkalemia — up to half of patients with K+ > 6.5
have no diagnostic changes. Treat the number and the patient, not the
tracing alone. But an abnormal tracing is an emergency, immediately.
Figure 31.5 — The electrocardiographic progression across the potassium range from 2.0 to 8.0 mEq/L, with the ionic mechanism of each change.
Described: A ladder of schematic electrocardiographic tracings arranged by serum potassium concentration from 8.0 down to 2.0 milliequivalents per litre, each paired with its mechanism. At 8.0 the tracing is a sine wave with no distinguishable P wave, QRS complex, or T wave, because sodium channels are largely inactivated and the QRS has merged with the T; this is a pre-arrest rhythm. At 7.5 the P wave has disappeared and the QRS is wide and slurred at more than 120 milliseconds with a tall peaked T, because atrial myocardium becomes inexcitable before ventricular myocardium and conduction slows throughout. At 6.5 the P wave is small and the PR interval long with a peaked T, as the depolarization threshold becomes harder to reach. At 5.5 the P wave and QRS are normal but the T wave is tall, narrow-based, symmetrical, and tented, because the conductance of the inward rectifier current is proportional to the square root of extracellular potassium, so repolarization is faster. At 4.0 the tracing is normal with a rounded T and no U wave, and the potassium equilibrium potential is minus 94 millivolts. At 3.0 the T wave flattens, the ST segment depresses, and a U wave appears after the T, because inward-rectifier conductance falls and repolarization is slowed. At 2.5 the U wave exceeds the T, which may be inverted or absent, the QU interval is long, and there is marked ST depression — the substrate for early afterdepolarizations and torsades de pointes. At 2.0 the T and U fuse and polymorphic ventricular tachycardia may appear, an effect greatly potentiated by digoxin at any low potassium. Amara at 2.9 is marked as showing a flat T wave, prominent U wave, prolonged QU interval, and occasional premature ventricular complexes. A caution states that the electrocardiogram is specific but not sensitive: up to half of patients with a potassium above 6.5 have no diagnostic changes, so a normal tracing does not exclude dangerous hyperkalemia, though an abnormal one is an immediate emergency.
Two further points about reading potassium on an ECG. First, the rate of change matters more than the absolute value. A patient whose potassium has risen from 4.0 to 6.5 over an hour is in far more danger than a dialysis patient who lives at 6.5, because the myocardium partially adapts to a chronically raised level. Second, peaked T waves are not specific: early repolarization, left ventricular hypertrophy, and hyperacute myocardial infarction all produce tall T waves. The hyperkalemic T is distinguished by being narrow-based, symmetrical, and tented — it looks as though someone pinched the apex — whereas the hyperacute T of infarction is broad-based and asymmetrical.
Exercise & Sport · Potassium, Sweat, and What Is Actually Lost
Two exercise phenomena in this section deserve real numbers.
Potassium leaves contracting muscle, and it matters. Every action potential in a muscle fiber loses a small amount of potassium to the interstitium during repolarization, and at high firing rates the Na⁺/K⁺ ATPase cannot keep up. Interstitial potassium in working muscle can reach 10–12 mEq/L, and arterial plasma potassium rises from 4.0 to 6–8 mEq/L during maximal exercise — a concentration that would be treated as an emergency in a hospital bed. Three things make it safe rather than lethal: it is brief (plasma potassium falls back within one to two minutes of stopping, as the pump catches up and muscle reclaims it); circulating catecholamines simultaneously drive potassium into cells through β₂ receptors, capping the rise; and the accompanying acidosis and temperature change alter cardiac excitability in partially offsetting directions. The rise is not merely tolerated — interstitial potassium is one of the signals that stimulates group III and IV muscle afferents, contributing to the exercise pressor reflex that raises heart rate and blood pressure to match the work. It also contributes to peripheral fatigue by depolarizing the fiber's T-tubule membrane and reducing action potential amplitude.
A practical consequence: a blood sample drawn immediately after hard exercise, or from a limb after prolonged fist clenching with a tourniquet in place, will read falsely high. That is one of the commonest causes of "pseudohyperkalemia" and of an unnecessary emergency call.
Sweat composition — what you actually lose.
| Component | Concentration in sweat | Plasma for comparison |
|---|---|---|
| Sodium | 20–80 mEq/L (mean ~35–45) | 140 |
| Chloride | 20–60 mEq/L | 104 |
| Potassium | 4–8 mEq/L | 4 |
| Magnesium, calcium | Trace | — |
| Osmolality | 80–200 mOsm/kg | 285 |
Sweat is always hypotonic to plasma, because the eccrine duct reabsorbs sodium and chloride from the primary secretion as it travels to the surface. That single fact has a large consequence: sweating loses proportionally more water than salt, so prolonged sweating without replacement raises plasma sodium and osmolality. Dehydration from exercise is hypertonic dehydration, and the correct response is water plus some salt — not salt tablets alone, and not enormous volumes of plain water (§31.9).
Worked, for Nia: a 3.5-hour marathon at a sweat rate of 1.2 L/h loses 4.2 L of sweat. At a sweat sodium of 45 mEq/L she loses 190 mEq of sodium — about 4.4 g of sodium, or 11 g of salt, which is roughly two to three days of a typical dietary intake. Against a total body sodium of about 3,300 mEq for her size, that is a loss of under 6%, which is why sodium depletion is rarely the primary problem — but it is more than enough to matter if she replaces the 4.2 L with plain water.
Acclimatization changes the numbers. After 7–14 days of heat exposure, aldosterone-driven ductal reabsorption can lower sweat sodium to as little as 5–10 mEq/L while sweat rate increases and sweating begins at a lower core temperature. The acclimatized athlete therefore sweats more, loses less salt per litre, and — importantly for §31.9 — produces an even more dilute sweat, which makes over-drinking relatively more dangerous rather than less.
Check Your Understanding 31.4
- A patient in diabetic ketoacidosis has a serum potassium of 5.4 mEq/L. Is her total body potassium high, normal, or low? What will happen when insulin is started, and what should be done first?
- Explain why hypokalemia prolongs the QT interval when the Nernst equation says a low extracellular potassium should hyperpolarize the cell and make it less excitable.
Show answers
- Her total body potassium is severely low — typically a deficit of 300 to 1,000 mEq — even though the serum value is high. Three mechanisms have moved potassium out of cells while the kidney was excreting it: insulin deficiency (the pump is under-stimulated), hypertonicity from hyperglycemia (water leaves cells, concentrating intracellular potassium and driving it down its gradient), and to a lesser extent the acidosis. Meanwhile the osmotic diuresis has been flushing potassium out of the body for hours to days, with the high distal flow and ketoanion delivery of Figure 31.4C accelerating it. Starting insulin reverses the shift and drives potassium rapidly back into cells; the serum value can fall precipitously, producing arrhythmia. The rule is therefore: check potassium before insulin, and if it is below 3.3 mEq/L, give potassium first and delay the insulin. At 5.4, insulin can be started but potassium must be added to the fluids once the level falls below about 5.0–5.3, with hourly monitoring.
- Because two different potassium currents are involved and they respond in opposite directions. The Nernst equation describes the equilibrium potential set by the potassium gradient, and a lower extracellular potassium does make that potential more negative — hence the hyperpolarization and the skeletal muscle weakness. But repolarization of the cardiac action potential depends on the conductance of potassium channels, and the conductance of the inward-rectifier current I_K1 is proportional to the square root of extracellular potassium. When extracellular potassium falls, that conductance falls, so potassium leaves the cell more slowly during phase 3, the action potential lengthens, and the QT (and the fused QU) interval prolongs. A long repolarization gives L-type calcium channels time to recover from inactivation and reopen, producing early afterdepolarizations, which in a heart with dispersed repolarization can trigger torsades de pointes. So hypokalemia makes the cell harder to excite and simultaneously harder to repolarize — and it is the repolarization defect that kills. This is a good demonstration that "excitability" is not one property, and that gradient and conductance are separate variables.
31.5 Calcium, Phosphate, and Magnesium
Calcium: total versus ionized
The body contains 1,000–1,200 g of calcium, 99% of it in bone as hydroxyapatite (Chapter 6). The circulating fraction is tiny and is defended within a very narrow band, because it controls neuromuscular excitability, excitation-contraction coupling, exocytosis of every neurotransmitter and hormone, and the coagulation cascade.
Plasma calcium exists in three fractions:
| Fraction | Proportion | Physiologically active? |
|---|---|---|
| Ionized (free) Ca²⁺ | ~45–50% | Yes — this is the regulated variable |
| Protein-bound (mostly albumin) | ~40–45% | No |
| Complexed (citrate, phosphate, bicarbonate, lactate) | ~10–15% | No |
Reference ranges: total calcium 8.5–10.5 mg/dL (2.12–2.62 mmol/L); ionized calcium 4.6–5.3 mg/dL (1.15–1.33 mmol/L).
Because most laboratories report total calcium while physiology responds to ionized calcium, two corrections matter.
Albumin correction. Roughly 0.8 mg/dL of total calcium is bound per 1 g/dL of albumin.
Corrected Ca (mg/dL) = measured Ca + 0.8 × (4.0 − albumin g/dL)
Worked: a malnourished patient, Ca 7.6 mg/dL, albumin 2.4 g/dL
corrected = 7.6 + 0.8 × (4.0 − 2.4) = 7.6 + 1.28 = 8.9 mg/dL
→ NORMAL. She does not have hypocalcemia; she has hypoalbuminemia.
pH correction — and this is the interesting one. Albumin carries many negatively charged sites, and H⁺ and Ca²⁺ compete for them.
- In alkalosis, H⁺ dissociates from albumin, exposing more negative binding sites, which then bind Ca²⁺. Ionized calcium falls while total calcium is unchanged.
- In acidosis, the reverse: H⁺ occupies the sites, calcium is displaced, and ionized calcium rises at an unchanged total.
Approximately, ionized calcium changes by 0.05 mmol/L (about 0.2 mg/dL) for every 0.1 unit change in pH, in the opposite direction.
This explains one of the classic bedside observations in all of medicine. A person having a panic attack hyperventilates to a pH of 7.55 — a rise of 0.15 — dropping ionized calcium by roughly 0.075 mmol/L. That is enough to lower the threshold for spontaneous firing of peripheral nerves, producing perioral and acral paresthesia, carpopedal spasm, and frank tetany, with a completely normal total serum calcium. Chvostek's and Trousseau's signs become positive. The treatment is to correct the ventilation, not to give calcium.
The mechanism is worth stating precisely, because "low calcium causes tetany" sounds backwards to students who remember that calcium is needed for contraction. Extracellular Ca²⁺ stabilizes the voltage-gated sodium channel: it screens negative surface charge on the membrane, so a larger depolarization is needed to open the channel. Lower the ionized calcium and the channel opens more easily — the neuron becomes hyperexcitable and fires spontaneously. Calcium's effect on excitability is opposite to its effect on contraction.
For Amara: her pH is 7.49, a rise of 0.09, so her ionized calcium is about 0.045 mmol/L below where it would be at pH 7.40. That is modest — enough to contribute to her cramps, not enough to cause tetany on its own. Combined with her hypokalemia and hypomagnesemia, though, it is the third of three simultaneous causes of neuromuscular irritability.
The regulatory loop
From Chapters 6 and 16, condensed and with the phosphate arm made explicit:
| Hormone | Trigger | Bone | Kidney | Gut | Net on Ca²⁺ | Net on PO₄³⁻ |
|---|---|---|---|---|---|---|
| PTH (chief cells) | ↓ ionized Ca²⁺ sensed by the calcium-sensing receptor (CaSR) | ↑ resorption (osteoblast RANKL → osteoclasts) | ↑ distal Ca²⁺ reabsorption (TRPV5); ↓ proximal PO₄³⁻ reabsorption (internalizes NaPi-IIa); ↑ 1α-hydroxylase | Indirect, via calcitriol | ↑ | ↓ |
| Calcitriol (1,25-(OH)₂ vitamin D) | PTH, low PO₄³⁻ | Permissive for resorption | ↑ Ca²⁺ and PO₄³⁻ reabsorption | ↑ Ca²⁺ absorption (TRPV6, calbindin) and PO₄³⁻ absorption | ↑ | ↑ |
| Calcitonin (parafollicular C cells) | ↑ Ca²⁺ | ↓ osteoclast activity | ↑ excretion | — | ↓ | ↓ |
| FGF23 (osteocytes) | ↑ PO₄³⁻, ↑ calcitriol | — | ↓ NaPi-IIa; ↓ 1α-hydroxylase | ↓ via less calcitriol | ↓ | ↓↓ |
The CaSR deserves a sentence of its own: it is a G-protein-coupled receptor whose ligand is an ion, not a peptide or a steroid. Extracellular calcium binds it directly on the parathyroid chief cell, and — unusually — activation inhibits PTH secretion. It is one of very few places in physiology where the regulated variable is sensed as itself, with no transduction step in between.
Two derived facts that examiners love and clinicians use daily:
- PTH raises calcium and lowers phosphate. Calcitriol raises both. So in primary hyperparathyroidism, calcium is high and phosphate low; in vitamin D toxicity, both are high. That single contrast separates most disorders of the axis.
- Calcium and phosphate precipitate together. When the Ca × PO₄ product exceeds roughly 55–70 mg²/dL², calcium phosphate deposits in soft tissue, vessels, and skin. This is why acute hyperphosphatemia — tumor lysis syndrome, a phosphate-containing enema in a patient with renal impairment, rhabdomyolysis — causes hypocalcemia: the calcium is precipitated out.
Phosphate
Serum phosphate 2.5–4.5 mg/dL; 85% of body phosphate is in bone, and most of the rest is intracellular as ATP, 2,3-BPG, creatine phosphate, nucleic acids, and membrane phospholipids. Serum phosphate is therefore, like potassium, a poor guide to total body content and highly sensitive to shifts.
Clinical Connection · Refeeding Syndrome — When Nutrition Is the Injury
A person who has eaten little for a week or more is in a catabolic, insulin-suppressed, glucagon-dominant state. Intracellular stores of phosphate, potassium, and magnesium have been quietly depleted — but their serum concentrations may be entirely normal, because the same catabolism that emptied the cells has been releasing these ions into the ECF, and the kidney has been excreting them. The laboratory sheet looks reassuring and the patient is dangerously depleted. This is the potassium lesson of §31.4, repeated for three ions at once.
Then feeding starts. Carbohydrate arrives; insulin surges; and three things happen within hours.
- Glucose is driven into cells and phosphorylated. Every glycolytic intermediate from glucose-6-phosphate onward consumes a phosphate group, and ATP regeneration consumes more. Serum phosphate can fall from normal to below 1.0 mg/dL within 24–72 hours.
- Potassium and magnesium follow glucose into cells, driven by the same insulin-stimulated Na⁺/K⁺ ATPase, and their serum levels crash in parallel.
- Thiamine is consumed as a cofactor for pyruvate dehydrogenase, and a depleted patient can be tipped into Wernicke encephalopathy by the feed itself.
The consequences of severe hypophosphatemia are all ATP-dependency failures: respiratory muscle weakness (sometimes ventilator dependence), reduced myocardial contractility and arrhythmia, hemolysis (the red cell needs ATP to maintain its membrane), rhabdomyolysis, seizures, and a fall in 2,3-BPG that left-shifts the oxyhemoglobin curve and impairs tissue oxygen unloading — a cruel addition, since oxygen delivery falls at the moment cellular oxygen demand rises.
Who is at risk: BMI under 16, unintentional weight loss over 15% in three to six months, negligible intake for more than 10 days, or low baseline potassium, phosphate, or magnesium. Also: prolonged alcohol use, anorexia nervosa, post-bariatric surgery, and prolonged ICU stays.
Management follows exactly from the mechanism. Start feeding at 10–20 kcal/kg/day — perhaps half of estimated needs — and advance slowly over 4–7 days. Give thiamine (typically 200–300 mg daily) before and with the first feed. Measure phosphate, potassium, and magnesium daily for the first week and replace aggressively; do not withhold nutrition to avoid the syndrome, but do not rush it.
The general principle is worth carrying beyond this box: an anabolic shift moves ions into cells. Refeeding, insulin therapy, treating megaloblastic anemia with B₁₂, and recovery from diabetic ketoacidosis all do the same thing, and all can drop serum potassium, phosphate, and magnesium sharply.
Magnesium — and the answer to Case File question 3
The body holds about 24 g (1,000 mmol) of magnesium: roughly 60% in bone, 20% in muscle, 20% in other soft tissue, and less than 1% in the extracellular fluid. Serum magnesium is 1.7–2.4 mg/dL (0.7–1.0 mmol/L) and, once again, is a poor index of total body content — a normal serum magnesium is entirely compatible with substantial depletion.
Magnesium is absorbed in the small intestine via TRPM6, and handled in the kidney in a way that is central to Amara's problem: the majority of filtered magnesium is reabsorbed paracellularly in the thick ascending limb, through the claudin-16/19 pore, driven by the lumen-positive transepithelial potential. Fine-tuning occurs in the distal convoluted tubule via TRPM6. Loop diuretics abolish the lumen-positive potential — that is mechanism 6 in §31.4 — and therefore waste magnesium obligatorily.
What magnesium does (an unusually long list for an ion):
- ATP is functionally Mg-ATP. Every kinase, every ATPase, and every polymerase requires magnesium as part of its substrate — including the Na⁺/K⁺ ATPase.
- Cofactor for more than 300 enzymes.
- Blocks the NMDA receptor channel pore at resting membrane potentials (Chapter 11).
- Required for PTH secretion and for normal end-organ responsiveness to PTH.
- Blocks the ROMK channel from the cytoplasmic side.
That last item is the one that answers question 3.
WHY HYPOKALEMIA IS REFRACTORY UNTIL MAGNESIUM IS REPLACED
NORMAL MAGNESIUM-DEPLETED
────────────────────────── ────────────────────────────
principal cell principal cell
ROMK channel ROMK channel
▓▓ ← intracellular Mg2+ ░░ ← Mg2+ block GONE
▓▓ plugs the pore from ░░
▓▓ the inside, LIMITING ░░ K+ flows out FREELY
││ K+ efflux ││ into the tubular lumen
▼▼ ▼▼▼▼▼
LUMEN — moderate K+ secretion LUMEN — UNRESTRAINED K+ WASTING
PLUS: Na+/K+ ATPase requires Mg-ATP → with low intracellular Mg2+ the pump
works less well → cells cannot RECAPTURE the K+ you administer
PLUS: hypomagnesemia mildly stimulates aldosterone secretion → lever 1 again
∴ ORAL OR IV POTASSIUM IS SECRETED STRAIGHT BACK OUT INTO THE URINE.
You can pour potassium in for days and the serum level will not move.
REPLACE MAGNESIUM FIRST, OR AT THE SAME TIME.
The same mechanism produces refractory hypocalcemia: magnesium depletion both impairs PTH secretion and blunts the target tissue response to PTH, so calcium will not correct until magnesium does. And magnesium depletion is independently arrhythmogenic — intravenous magnesium is the treatment for torsades de pointes regardless of the measured serum magnesium, because it suppresses the early afterdepolarizations that trigger it.
Hypermagnesemia is almost always iatrogenic or a consequence of renal failure, and it progresses in a stereotyped order that makes bedside monitoring possible: loss of deep tendon reflexes at roughly 4–6 mg/dL, then hypotension and somnolence, then respiratory depression around 10–12 mg/dL, then cardiac arrest above about 15 mg/dL. This ordering is why the deep tendon reflex check is the standard bedside monitor during magnesium infusion for preeclampsia — the therapy Nia's obstetric team will have discussed with her (Chapter 28).
Check Your Understanding 31.5
- A patient with a severe head injury is hyperventilated to a PaCO₂ of 25 mm Hg to reduce intracranial pressure. Her total serum calcium is 9.4 mg/dL — normal. She develops carpopedal spasm. Explain, and say what you would measure.
- A patient on long-term furosemide has a potassium of 2.8 mEq/L that has not responded to three days of 80 mEq/day of oral potassium chloride. What single test would you order, and what would you predict?
Show answers
- Acute respiratory alkalosis has raised her pH, and H⁺ has therefore dissociated from albumin, exposing negatively charged binding sites that now bind calcium. Her ionized calcium has fallen — roughly 0.05 mmol/L for each 0.1 unit of pH rise, so a pH of about 7.55 from a baseline 7.40 would drop it by around 0.075 mmol/L — while her total calcium is completely unchanged, because no calcium has left the body. Lower ionized calcium destabilizes voltage-gated sodium channels by reducing membrane surface-charge screening, so peripheral nerves become hyperexcitable and fire spontaneously: paresthesia, carpopedal spasm, positive Chvostek and Trousseau signs. Measure ionized calcium, not total, and simultaneously check the arterial blood gas. The correct treatment is to reduce the degree of hyperventilation — not to give calcium — since the total body calcium is normal. This case also illustrates why prophylactic hyperventilation for raised intracranial pressure has fallen out of favor: the cerebral vasoconstriction that lowers pressure also lowers cerebral blood flow, and the metabolic side effects are real.
- Serum magnesium, and I would predict it to be low. Furosemide abolishes the lumen-positive transepithelial potential of the thick ascending limb, which is the driving force for paracellular magnesium reabsorption, so magnesium is wasted obligatorily alongside potassium. Intracellular magnesium normally blocks the ROMK channel from the cytoplasmic side and limits potassium secretion; when intracellular magnesium falls, that block is relieved and ROMK conducts potassium into the tubular lumen without restraint. In addition, the Na⁺/K⁺ ATPase requires Mg-ATP, so cells cannot efficiently take up the potassium being administered. The result is that oral or intravenous potassium is secreted straight back into the urine. Replacing magnesium — and continuing to replace it, since a normal serum level does not exclude total body depletion — is what allows the potassium to correct. Note that the same mechanism explains refractory hypocalcemia in magnesium depletion, through impaired PTH secretion and PTH resistance.
31.6 Acid-Base: The Fundamentals
What pH actually means, in clinical numbers
pH is the negative base-10 logarithm of the hydrogen ion concentration. Two consequences of that definition get lost in the notation and both matter.
First, the concentrations involved are vanishingly small.
| pH | [H⁺] | Comment |
|---|---|---|
| 6.80 | 158 nmol/L | Approximate lower limit of survival |
| 7.00 | 100 nmol/L | Severe acidemia |
| 7.20 | 63 nmol/L | Significant acidemia |
| 7.40 | 40 nmol/L | Normal |
| 7.50 | 32 nmol/L | Alkalemia |
| 7.70 | 20 nmol/L | Severe alkalemia |
| 7.80 | 16 nmol/L | Approximate upper limit of survival |
Compare 40 nanomoles per litre of H⁺ with 140 millimoles per litre of sodium: a difference of a factor of about 3.5 million. The body regulates hydrogen ion concentration roughly a million times more precisely, in absolute terms, than it regulates its most abundant cation — and it does so because proteins are exquisitely sensitive to it.
Second, the scale is logarithmic, so equal pH steps are not equal chemical steps. Going from pH 7.4 to 7.1 doubles [H⁺]; going from 7.4 to 7.7 halves it. A useful bedside approximation for the clinical range 7.25–7.50:
[H+] in nmol/L ≈ 80 − (the last two digits of the pH)
pH 7.40 → 80 − 40 = 40 nmol/L (true value 40) ✔
pH 7.30 → 80 − 30 = 50 nmol/L (true value 50) ✔
pH 7.49 → 80 − 49 = 31 nmol/L (true value 32) ✔
Why proteins care. Hydrogen ions bind to and dissociate from the ionizable side chains of proteins — above all the imidazole group of histidine, whose pKa of about 6.5–7.0 sits close to physiological pH, making it the most pH-sensitive residue in the proteome. Changing the charge on a residue changes the electrostatic forces holding the protein's tertiary structure, and therefore its shape and its function. The clinical consequences follow directly:
| pH | Effects |
|---|---|
| < 7.20 | Myocardial contractility falls; vascular smooth muscle relaxes (hypotension); responsiveness to catecholamines falls; arrhythmia threshold falls; hyperkalemia in mineral acidoses; insulin resistance |
| < 7.10 | The above become severe; catecholamine infusions lose effect — which is why acidosis must be corrected for pressors to work |
| > 7.55 | Cerebral vasoconstriction (confusion, seizures); ionized hypocalcemia and tetany; hypokalemia; arrhythmia; the oxyhemoglobin curve left-shifts, impairing tissue O₂ unloading |
The two acid loads
| Volatile acid | Fixed (nonvolatile) acid | |
|---|---|---|
| Chemical identity | CO₂ (as carbonic acid) | Sulfuric acid, phosphoric acid, organic acids |
| Source | Oxidative metabolism of all fuels | Sulfur amino acids (methionine, cysteine); phospholipids and phosphoproteins; incompletely metabolized organics |
| Daily quantity | ~15,000 mmol/day | ~70 mEq/day (range 50–100; ~1 mEq/kg) |
| Route of removal | Lungs, by exhalation | Kidneys, by excretion |
| Time to eliminate | Minutes | Hours to days |
The ratio is arresting: the lungs handle roughly 200 times more acid per day than the kidneys. And yet the kidney's contribution is irreplaceable, because fixed acid cannot be exhaled. If ventilation stops, PaCO₂ rises and pH falls within minutes; if the kidneys stop, fixed acid accumulates at 70 mEq/day and pH falls over days. Different mechanisms, different timescales, both lethal.
The chemistry that links CO₂ to acid is one reaction, and it is worth writing out:
CO2 + H2O ⇌ᶜᵃ H2CO3 ⇌ H+ + HCO3-
▲
└── carbonic anhydrase: one of the fastest enzymes known,
~10^6 reactions per second. Present in red cells,
renal tubular cells, gastric parietal cells, and
(as CA IV) on the luminal surface of the proximal tubule.
The reaction is freely reversible, which is why CO₂ is simultaneously a waste gas and an acid, and why the lungs are an acid-base organ.
The three lines of defense
TIME AXIS (LOGARITHMIC)
│
│ 1 sec 10 sec 1 min 10 min 1 h 10 h 1 day 3 d
├──────────┬──────────┬─────────┬─────────┬─────────┬────────┬───────┬────►
│ │ │ │ │ │ │ │
│▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓│ │ │ │ │ │
│ 1 · CHEMICAL BUFFERS │ │ │ │ │ │
│ INSTANT — but they only CONVERT a strong acid into a weak one. │
│ Nothing leaves the body. They BUY TIME. │
│ · BICARBONATE / CO2 ~50% of whole-blood buffering, ALL of the │
│ extracellular buffering that matters — and the only OPEN │
│ system (see below) │
│ · HEMOGLOBIN the largest single protein buffer; handles │
│ most of the H+ generated by CO2 hydration in the red cell │
│ · PLASMA PROTEINS ~7% (histidine imidazole groups) │
│ · PHOSPHATE minor in ECF (1 mmol/L) but MAJOR inside │
│ cells (75+ mmol/L) and in URINE (titratable acid, §31.7) │
│ · INTRACELLULAR PROTEIN — cells absorb ~50% of an acid load in │
│ hours by exchanging H+ IN for K+ OUT ← this is why acidosis │
│ raises serum potassium │
│ · BONE CARBONATE a slow, large reservoir; consumed in │
│ chronic acidosis — one reason CKD causes bone disease │
│ │ │ │ │ │ │ │
│ │ ▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓│ │ │
│ │ 2 · RESPIRATORY COMPENSATION │ │ │
│ │ starts in 1–2 MIN · maximal by 12–24 H │
│ │ chemoreceptors → ventilation → PaCO2 │
│ │ DOUBLE alveolar ventilation → HALVE PaCO2 │
│ │ Corrects 50–75% of a metabolic disturbance. │
│ │ NEVER fully — full correction would remove the │
│ │ stimulus driving it. │
│ │ │ │ │ │ │ │
│ │ │ │ ▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓
│ │ │ │ 3 · RENAL COMPENSATION │
│ │ │ │ starts within HOURS │
│ │ │ │ maximal at 3–5 DAYS │
│ │ │ │ · reabsorb filtered HCO3- │
│ │ │ │ · generate NEW HCO3- │
│ │ │ │ · excrete acid as │
│ │ │ │ TITRATABLE ACID + NH4+ │
│ │ │ │ THE ONLY DEFINITIVE ROUTE: │
│ │ │ │ the only one that removes │
│ │ │ │ acid FROM THE BODY │
├──────────┴──────────┴─────────┴─────────┴─────────┴────────┴───────┴────►
│ 1 sec 10 sec 1 min 10 min 1 h 10 h 1 day 3 d
★ THE CLINICAL PAYOFF OF THE TIME AXIS:
A raised HCO3- in a patient with a raised PaCO2 tells you the problem has
been present for DAYS, not hours — because renal compensation is slow.
The TIME COURSE IS DIAGNOSTIC. §31.8 turns this into a formula.
Figure 31.6 — The three lines of acid-base defense plotted on a logarithmic time axis, from buffering in seconds to renal compensation over days.
Described: A logarithmic time axis running from one second to three days, with three overlapping response bands. The first band, chemical buffers, is active essentially instantly and spans the first minute; the caption notes that buffers only convert a strong acid into a weak one, remove nothing from the body, and therefore buy time. Its components are listed: bicarbonate and carbon dioxide, providing about half of whole-blood buffering and the only open buffer system; hemoglobin, the largest single protein buffer, handling most of the hydrogen ions generated by carbon dioxide hydration inside the red cell; plasma proteins at about seven percent through histidine imidazole groups; phosphate, minor in extracellular fluid at one millimole per litre but major inside cells at over 75 and in urine as titratable acid; intracellular protein, with cells absorbing about half of an acid load over hours by exchanging hydrogen ions inward for potassium ions outward, which is why acidosis raises serum potassium; and bone carbonate, a slow, large reservoir consumed in chronic acidosis and one reason chronic kidney disease causes bone disease. The second band, respiratory compensation, begins at one to two minutes and is maximal by twelve to twenty-four hours: chemoreceptors adjust ventilation and hence arterial carbon dioxide tension, with a doubling of alveolar ventilation halving the carbon dioxide tension. It corrects fifty to seventy-five percent of a metabolic disturbance but never fully, because full correction would remove the stimulus driving it. The third band, renal compensation, begins within hours and is maximal at three to five days, and comprises reabsorption of filtered bicarbonate, generation of new bicarbonate, and excretion of acid as titratable acid and ammonium. It is described as the only definitive route because it is the only one that removes acid from the body. A closing note gives the clinical payoff: a raised bicarbonate accompanying a raised carbon dioxide tension indicates a problem present for days rather than hours, because renal compensation is slow, so the time course is itself diagnostic.
The bicarbonate buffer system — plain language first
Here is the whole thing without any mathematics.
Blood pH depends on the RATIO of bicarbonate to dissolved carbon dioxide — not on either one alone. Normally that ratio is about 20 : 1. Keep the ratio at 20 : 1 and the pH is 7.40, whatever the absolute numbers happen to be. And because two different organs control the two numbers — the lungs set the CO₂, the kidneys set the bicarbonate — either organ can adjust the ratio to compensate for a problem in the other.
That paragraph is the entire conceptual content of acid-base physiology. Everything that follows is arithmetic and detail.
Now the mathematics, twice.
Henderson-Hasselbalch, the version everyone quotes:
pH = 6.1 + log ( [HCO3-] / (0.03 × PaCO2) )
6.1 = pKa of the carbonic acid system
0.03 = solubility coefficient of CO2 (mmol/L per mm Hg)
NORMAL: pH = 6.1 + log ( 24 / (0.03 × 40) )
= 6.1 + log ( 24 / 1.2 )
= 6.1 + log 20
= 6.1 + 1.30 = 7.40 ← there is the 20 : 1 ratio
The Henderson equation, which is easier arithmetic and far more useful at the bedside:
[H+] (nmol/L) = 24 × PaCO2 / [HCO3-]
NORMAL: 24 × 40 / 24 = 40 nmol/L → pH 7.40 ✔
AMARA: 24 × 47 / 34 = 33.2 nmol/L
→ pH ≈ 7.48, which matches her reported 7.49 within
measurement error. HER BLOOD GAS IS INTERNALLY CONSISTENT.
Use the Henderson equation to check every blood gas you are handed. The three values — pH, PaCO₂, and bicarbonate — are not independent; the bicarbonate is usually calculated from the other two. If they do not agree, the sample is a venous gas mislabelled as arterial, a delayed or air-contaminated specimen, or a transcription error, and interpreting it will lead you somewhere wrong.
╔════════════════════════════════════════════════════════════════════════╗
║ THE BICARBONATE BUFFER SYSTEM AS A TWO-ORGAN MACHINE ║
╚════════════════════════════════════════════════════════════════════════╝
┌──────────────────────┐ ┌──────────────────────┐
│ LUNGS │ │ KIDNEYS │
│ set the DENOMINATOR │ │ set the NUMERATOR │
│ │ │ │
│ PaCO2 35–45 mm Hg │ │ HCO3- 22–26 mEq/L │
│ response: MINUTES │ │ response: HOURS–DAYS│
└──────────┬───────────┘ └──────────┬───────────┘
│ │
│ ↑ ventilation → ↓ PaCO2 │ ↑ H+ excretion
│ ↓ ventilation → ↑ PaCO2 │ → ↑ HCO3-
│ │ ↑ HCO3- excretion
▼ ▼ → ↓ HCO3-
┌────────────────────────────────────────────────────────────────────┐
│ │
│ [ HCO3- ] 24 │
│ pH ∝ ────────────────── = ──────── = 20 : 1 │
│ 0.03 × [ PaCO2 ] 1.2 │
│ │
│ KEEP THE RATIO AT 20 : 1 AND pH = 7.40 │
│ — regardless of the absolute numbers. │
└────────────────────────────────────────────────────────────────────┘
▲ ▲
│ │
┌───────────────┴──────────────┐ ┌──────────────────┴──────────────┐
│ CO2 + H2O ⇌ H2CO3 ⇌ H+ + HCO3-│ │ FIXED ACID (~70 mEq/day) │
│ (carbonic anhydrase) │ │ H2SO4 from methionine/cysteine │
│ ~15,000 mmol/day of CO2 │ │ H3PO4 from phospholipids │
│ from oxidative metabolism │ │ organic acids │
└───────────────────────────────┘ └─────────────────────────────────┘
╔══ WHY THIS BUFFER WORKS DESPITE A "WRONG" pKa ═══════════════════════╗
║ A buffer works best within ±1 pH unit of its pKa. This system's ║
║ pKa is 6.1 — a full 1.3 units from 7.4. On paper it should be a ║
║ poor buffer. It is in fact the best one the body has, for two ║
║ reasons that no closed buffer can match: ║
║ ║
║ 1 · IT IS AN OPEN SYSTEM. The acid member of the pair (CO2) is ║
║ continuously BLOWN OFF by the lungs, so it never accumulates ║
║ and the reaction is dragged to the right indefinitely. ║
║ 2 · BOTH MEMBERS ARE INDEPENDENTLY REGULATED, by two different ║
║ organs, on two different time scales. ║
║ ║
║ A closed buffer with the same pKa would be nearly useless. The ║
║ physiology, not the chemistry, is what makes it work. ║
╚══════════════════════════════════════════════════════════════════════╝
╔══ HEMOGLOBIN AND THE ISOHYDRIC SHIFT ════════════════════════════════╗
║ In the TISSUES: Hb releases O2 → DEOXYhemoglobin, which is a ║
║ BETTER proton acceptor (higher pKa of its histidines) → it ║
║ absorbs exactly the H+ that CO2 hydration is producing there. ║
║ In the LUNGS: Hb binds O2 → OXYhemoglobin releases those H+ → ║
║ they recombine with HCO3- → CO2 → exhaled. ║
║ The molecule that carries oxygen is the same molecule that buffers ║
║ the acid generated by using it, and the coupling is automatic. ║
║ (The Haldane effect, Chapter 22, seen from the acid-base side.) ║
╚══════════════════════════════════════════════════════════════════════╝
Figure 31.7 — The bicarbonate buffer system as a two-organ machine: lungs control the carbon dioxide, kidneys control the bicarbonate, and pH depends on the ratio.
Described: A diagram with the lungs on the left, setting the denominator of the ratio by controlling arterial carbon dioxide tension between 35 and 45 mm Hg with a response time of minutes, and the kidneys on the right, setting the numerator by controlling bicarbonate between 22 and 26 milliequivalents per litre with a response time of hours to days. Increasing ventilation lowers carbon dioxide and decreasing it raises carbon dioxide; increasing renal hydrogen ion excretion raises bicarbonate while increasing bicarbonate excretion lowers it. Both feed into a central box showing that pH is proportional to bicarbonate divided by 0.03 times carbon dioxide tension, normally 24 over 1.2, a ratio of 20 to 1, and that keeping the ratio at 20 to 1 keeps pH at 7.40 regardless of the absolute numbers. Beneath, two source boxes: carbon dioxide plus water in equilibrium with carbonic acid and then with hydrogen and bicarbonate ions, catalyzed by carbonic anhydrase, generating about 15,000 millimoles per day from oxidative metabolism; and fixed acid at about 70 milliequivalents per day, comprising sulfuric acid from methionine and cysteine, phosphoric acid from phospholipids, and organic acids. A boxed note explains why the system works despite a pKa of 6.1, a full 1.3 units from physiological pH: it is an open system in which carbon dioxide is continuously blown off by the lungs so the acid member never accumulates, and both members are independently regulated by two different organs on two different timescales — a closed buffer with the same pKa would be nearly useless. A final boxed note describes the isohydric shift: in the tissues, hemoglobin releasing oxygen becomes deoxyhemoglobin, a better proton acceptor, absorbing exactly the hydrogen ions that carbon dioxide hydration is generating there; in the lungs, hemoglobin binding oxygen releases those hydrogen ions, which recombine with bicarbonate to form carbon dioxide for exhalation. The molecule that carries oxygen is the same molecule that buffers the acid generated by using it.
Exercise & Sport · The Acid-Base Response to Maximal Exercise
Hard exercise is the largest acid load a healthy body ever generates voluntarily, and watching the three lines of defense meet it is the best available demonstration of Figure 31.6.
What happens, in order. At low intensity, ATP is supplied aerobically and blood lactate stays near 1 mmol/L. As intensity rises past roughly 50–60% of VO₂max, glycolytic flux exceeds the capacity of pyruvate to enter mitochondria, and lactate production exceeds clearance. Blood lactate rises steeply — to 15–25 mmol/L at exhaustion in a trained athlete.
A note on chemistry, because the physiology is often taught wrongly: lactate itself is not the acid. At physiological pH lactic acid is essentially fully dissociated, and the protons that acidify the muscle come principally from the net hydrolysis of ATP when glycolytic ATP supply exceeds mitochondrial ATP resynthesis. Lactate production actually consumes a proton. Lactate is the best available marker of the acidosis, not its cause — and it is also a fuel, shuttled to the heart, to oxidative muscle fibers, and to the liver for gluconeogenesis (the Cori cycle, Chapter 24).
The numbers at exhaustion:
| Variable | Rest | Maximal exercise |
|---|---|---|
| Arterial pH | 7.40 | 7.10–7.25 |
| Intramuscular pH | ~7.0 | 6.4–6.6 |
| Arterial HCO₃⁻ | 24 mEq/L | 10–14 mEq/L |
| PaCO₂ | 40 mm Hg | 25–30 mm Hg |
| Blood lactate | ~1 mmol/L | 15–25 mmol/L |
| Plasma K⁺ | 4.0 mEq/L | 6–8 mEq/L |
| Minute ventilation | 6 L/min | 150–200 L/min |
All three lines of defense, visible in one graph. Bicarbonate is consumed almost immediately — it is the buffer being spent, and its fall from 24 to 12 represents roughly 12 mEq/L of acid absorbed across the whole extracellular volume. Intramuscular buffering is handled by phosphate, by carnosine (a dipeptide whose histidine has a pKa near 7.0, which is why beta-alanine supplementation, which raises muscle carnosine, has a measurable ergogenic effect in 1–4 minute events), and by protein. Respiratory compensation appears as the classic ventilatory threshold: ventilation, which had been rising in proportion to CO₂ production, suddenly rises faster as extra CO₂ is generated from bicarbonate buffering the metabolic acid. For a period, that hyperventilation exactly offsets the acid load and arterial pH is held constant — the isocapnic buffering phase. Beyond the respiratory compensation point, ventilation cannot keep up and pH falls steeply.
Renal compensation contributes essentially nothing, because the entire episode is over in minutes and the kidney's response takes hours. Recovery is instead by metabolic clearance: lactate is oxidized or converted back to glucose, consuming protons, and bicarbonate is regenerated within 30–60 minutes.
Two applications. First, this is why sodium bicarbonate loading is a real (if gastrointestinally unpleasant) ergogenic aid for events of 1–7 minutes: raising extracellular bicarbonate increases the gradient for proton efflux from muscle. Second, it explains why an athlete's post-exercise blood gas can look alarming — pH 7.15, bicarbonate 12, potassium 6.5 — in a person who is entirely healthy and will be normal within an hour. Context determines meaning; numbers alone do not.
31.7 Respiratory and Renal Control of pH
The lung, in minutes
Ventilation controls PaCO₂ and therefore the denominator of the ratio. The relationship is inverse and near-exact: alveolar ventilation and PaCO₂ are inversely proportional at a fixed CO₂ production. Double alveolar ventilation and PaCO₂ halves; halve it and PaCO₂ doubles.
The controller (Chapter 22):
- Central chemoreceptors in the ventral medulla sense the [H⁺] of cerebrospinal fluid. This detail matters: CO₂ crosses the blood-brain barrier freely while H⁺ and HCO₃⁻ do not. So a rise in arterial PaCO₂ is transmitted immediately into the CSF, where carbonic anhydrase converts it to H⁺ and the receptors respond within seconds. A metabolic acidosis, by contrast, reaches the central chemoreceptors only slowly and indirectly, which is why its respiratory compensation takes 12–24 hours to become maximal.
- Peripheral chemoreceptors in the carotid and aortic bodies respond to PaO₂ below about 60 mm Hg, to PaCO₂, and — importantly here — directly to arterial [H⁺]. They are the route by which a metabolic acidosis produces immediate hyperventilation.
In severe metabolic acidosis, the pattern is Kussmaul respiration: deep and rapid. The depth matters more than the rate, because increasing tidal volume raises alveolar ventilation more efficiently than increasing frequency, which mostly moves dead space air back and forth.
The kidney, in hours to days
The kidney has three jobs, and students routinely collapse them into one. They are separate.
Job 1 — Reabsorb the filtered bicarbonate. (Losing it is equivalent to adding acid.)
At a GFR of 180 L/day and a plasma bicarbonate of 24 mEq/L, roughly 4,300 mEq of bicarbonate is filtered every day — sixty times the entire daily fixed acid load. Losing even a few percent would be catastrophic, so essentially all of it is reclaimed: 80–90% in the proximal tubule, ~10% in the thick ascending limb, and the remainder distally.
Note carefully: this job generates no new bicarbonate. The H⁺ secreted into the lumen is consumed in the process and simply returns what was filtered.
Job 2 — Generate new bicarbonate by excreting titratable acid. Roughly 10–40 mEq/day.
Job 3 — Generate new bicarbonate by excreting ammonium. Roughly 30–40 mEq/day at baseline, and — this is the crucial point — adaptable up to 300–400 mEq/day over several days. This is the regulated arm, and the reason renal compensation takes days rather than hours: glutaminase and the ammoniagenic machinery must be induced.
╔══ JOB 1 · RECLAIM FILTERED HCO3- ═══ PROXIMAL TUBULE CELL ═════════════╗
║ LUMEN CELL BLOOD ║
║ ║
║ filtered HCO3- ─────┐ ║
║ + │ ║
║ H+ ◄── NHE3 ◄──┼── H+ ◄─┐ ║
║ │ ▲ │ │ ║
║ │ └─Na+──┼──► Na+ │ CO2 + H2O ║
║ ▼ │ │ │ carbonic anhydrase II ║
║ H2CO3 │ └── H+ ◄┴─► HCO3- ──► NBCe1 ──► HCO3- ║
║ │ CA IV │ (Na+/3HCO3-) to ║
║ ▼ (luminal │ blood ║
║ CO2 + H2O ─────────┼──► CO2 diffuses IN ║
║ surface) │ ║
║ ★ NET RESULT: the filtered HCO3- is RETURNED to the blood. ║
║ NO NEW BICARBONATE IS MADE. The H+ is recycled, not excreted. ║
║ (Acetazolamide blocks CA here → bicarbonate wasting → the ║
║ therapeutic non-gap acidosis used in altitude sickness.) ║
╚════════════════════════════════════════════════════════════════════════╝
╔══ JOB 2 · TITRATABLE ACID ═══ COLLECTING DUCT α-INTERCALATED CELL ═════╗
║ LUMEN CELL BLOOD ║
║ ║
║ filtered HPO4²⁻ CO2 + H2O ║
║ + │ CA II ║
║ H+ ◄── H+-ATPase ◄─────────────┤ ║
║ H+ ◄── H+/K+-ATPase ◄──────────┤ ║
║ │ (also reclaims K+)│ ║
║ ▼ HCO3- ──► AE1 (Cl-/HCO3-) ──► HCO3- ║
║ H2PO4⁻ EXCRETED ▲ to ║
║ └── Cl- ◄── blood ║
║ ★ Because the H+ leaves attached to PHOSPHATE, a NEW HCO3- can be ║
║ exported to the blood. ~10–40 mEq/day. ║
║ ★ LIMITED BY: filtered phosphate load, and a MINIMUM URINE pH of ║
║ ~4.4 (below which the pump cannot work against the gradient). ║
╚════════════════════════════════════════════════════════════════════════╝
╔══ JOB 3 · AMMONIUM — THE ADAPTABLE, HIGH-CAPACITY ROUTE ═══════════════╗
║ ║
║ PROXIMAL TUBULE: GLUTAMINE ──glutaminase──► 2 NH4+ + 2 HCO3- ║
║ (the α-ketoglutarate → glucose or CO2 → HCO3-) ║
║ │ │ ║
║ │ NH4+ secreted into └──► NEW HCO3- ║
║ │ lumen (rides on NHE3 to BLOOD ║
║ ▼ in place of H+) ║
║ THICK ASC. LIMB: NH4+ REABSORBED (rides on NKCC2 in place of K+) ║
║ │ ║
║ ▼ accumulates in the MEDULLARY INTERSTITIUM ║
║ COLLECTING DUCT: NH3 diffuses into the lumen (lipid-soluble) ║
║ + H+ pumped in by the H+-ATPase ║
║ = NH4+ — now CHARGED, so it CANNOT diffuse ║
║ back out. "DIFFUSION TRAPPING." ║
║ ║
║ ★ EACH NH4+ EXCRETED = ONE NEW HCO3- ADDED TO THE BLOOD. ║
║ ★ 30–40 mEq/day baseline → 300–400 mEq/day in chronic acidosis. ║
║ Induction takes 3–5 DAYS. THAT is why renal compensation is slow. ║
╚════════════════════════════════════════════════════════════════════════╝
NET ACID EXCRETION = TITRATABLE ACID + NH4+ − urinary HCO3-
≈ 70 mEq/day, matching the fixed acid load exactly
★ WHY FREE H+ IS IRRELEVANT: at the minimum urine pH of 4.4, [H+] is
40 µmol/L. In 1.5 L of urine that is 0.06 mEq — one-thousandth of the
daily load. ESSENTIALLY ALL ACID LEAVES THE BODY BOUND TO A BUFFER.
Figure 31.8 — The three renal jobs in acid-base balance, drawn as tubular cells: reclaiming filtered bicarbonate, generating new bicarbonate with phosphate, and generating new bicarbonate with ammonia.
Described: Three tubular cell diagrams. The first shows a proximal tubule cell reclaiming filtered bicarbonate: hydrogen ions are secreted into the lumen by the sodium-hydrogen exchanger NHE3 in exchange for sodium entering the cell; in the lumen the hydrogen ion combines with filtered bicarbonate to form carbonic acid, which luminal carbonic anhydrase IV converts to carbon dioxide and water; the carbon dioxide diffuses into the cell where intracellular carbonic anhydrase II regenerates hydrogen ion and bicarbonate; the bicarbonate leaves across the basolateral membrane on the sodium-three-bicarbonate cotransporter NBCe1 into the blood. The net result is that filtered bicarbonate is returned to the blood, no new bicarbonate is made, and the hydrogen ion is recycled rather than excreted; acetazolamide blocks carbonic anhydrase here, producing the therapeutic bicarbonate-wasting acidosis used in altitude sickness. The second diagram shows a collecting duct type A intercalated cell excreting titratable acid: hydrogen ions are pumped into the lumen by an apical hydrogen ATPase and a hydrogen-potassium ATPase, where they combine with filtered monohydrogen phosphate to form dihydrogen phosphate, which is excreted; because the hydrogen ion leaves attached to phosphate, a new bicarbonate generated inside the cell can be exported to blood on the basolateral chloride-bicarbonate exchanger AE1. This route provides 10 to 40 milliequivalents per day and is limited by the filtered phosphate load and by a minimum urine pH near 4.4. The third diagram shows ammonium excretion: proximal tubule cells metabolize glutamine by glutaminase into two ammonium ions and two bicarbonate ions; the ammonium is secreted into the lumen on NHE3 in place of hydrogen, reabsorbed in the thick ascending limb on NKCC2 in place of potassium, accumulated in the medullary interstitium, and finally trapped in the collecting duct lumen, where lipid-soluble ammonia diffuses in, meets pumped hydrogen ions, and becomes charged ammonium that cannot diffuse back — diffusion trapping. Each ammonium excreted adds one new bicarbonate to the blood, and the route expands from 30 to 40 milliequivalents per day at baseline to 300 to 400 in chronic acidosis, with induction taking three to five days, which is why renal compensation is slow. A closing note gives net acid excretion as titratable acid plus ammonium minus urinary bicarbonate, about 70 milliequivalents per day, matching the fixed acid load, and observes that free hydrogen ion excretion is negligible: at the minimum urine pH of 4.4, 1.5 litres of urine carries only 0.06 milliequivalents, one-thousandth of the daily load, so essentially all acid leaves the body bound to a buffer.
Histology · The Collecting Duct's Two Cell Types, and the Cell That Can Reverse Itself
Down a microscope, the collecting duct epithelium is obviously composed of two different cells, and the difference in appearance is a direct readout of the difference in job.
Principal cells (about two-thirds of the population) are pale-staining, with relatively few mitochondria, sparse short microvilli, and a single central primary cilium projecting into the lumen. They look unremarkable because their work is largely done by channels moving ions down electrochemical gradients rather than by pumps.
- Apical: ENaC (sodium in), ROMK and BK (potassium out), aquaporin-2 (ADH-regulated water channels).
- Basolateral: Na⁺/K⁺ ATPase, aquaporins 3 and 4.
- Jobs: sodium reabsorption, potassium secretion, water reabsorption. The targets of aldosterone and of ADH.
- That single primary cilium is not decorative: it bends with tubular flow, raising intracellular calcium and opening the BK channels — the flow sensor of lever 3 in §31.4.
Intercalated cells (about one-third) are dark-staining, packed with mitochondria, and carry abundant apical microplicae and microvilli. They look like cells doing heavy pumping work, because they are: primary active transport of protons against a gradient of up to a thousandfold is metabolically expensive.
| Type A (α) intercalated cell | Type B (β) intercalated cell | |
|---|---|---|
| Job | Secretes acid | Secretes base |
| Apical membrane | H⁺-ATPase; H⁺/K⁺-ATPase | Pendrin (Cl⁻/HCO₃⁻ exchanger) |
| Basolateral membrane | AE1 (band 3, Cl⁻/HCO₃⁻ exchanger) — the same protein that gives the red cell its chloride shift | H⁺-ATPase |
| Net effect | H⁺ into urine, new HCO₃⁻ into blood | HCO₃⁻ into urine, H⁺ into blood |
| Upregulated in | Acidosis | Alkalosis |
The two are mirror images: type B is type A with the apical and basolateral machinery swapped. And a third population, "non-A non-B," carries pendrin and the H⁺-ATPase on the apical membrane.
The remarkable part is plasticity. The ratio of type A to type B cells shifts with chronic acid-base status, and there is good evidence that individual cells can interconvert — remodeling their membranes and redistributing transporters — under the influence of the extracellular matrix protein hensin. A single epithelium can reverse the direction of its net transport, which is an unusual thing for a differentiated cell to do and a striking piece of structure-follows-function.
One clinical payoff. Pendrin exchanges luminal chloride for bicarbonate. If there is no chloride in the lumen — because the patient is chloride-depleted from vomiting or a loop diuretic — pendrin cannot operate, and the kidney becomes unable to excrete the excess bicarbonate. That is the molecular reason a metabolic alkalosis is called "chloride-responsive," why it persists indefinitely without chloride replacement, and why Amara's alkalosis will not resolve until she is given chloride — best given as potassium chloride, which corrects two of her abnormalities at once.
Check Your Understanding 31.7
- Why does respiratory compensation for a metabolic acidosis take 12–24 hours to reach its maximum, when the respiratory system can change PaCO₂ within a minute?
- A patient has a metabolic acidosis with a urine pH of 5.9 and a positive urine anion gap. What does this tell you, and how does it differ from a patient with diarrhea?
Show answers
- Because the central chemoreceptors, which provide most of the ventilatory drive, sense the hydrogen ion concentration of cerebrospinal fluid, and the blood-brain barrier is freely permeable to CO₂ but not to H⁺ or HCO₃⁻. In a metabolic acidosis the arterial [H⁺] rises but CSF [H⁺] does not change immediately; the central receptors are effectively shielded. The immediate hyperventilation therefore comes from the peripheral chemoreceptors, which do respond directly to arterial [H⁺] but contribute a minority of the total drive. Over the following hours, bicarbonate is slowly transported out of the CSF, CSF pH falls to match, and the central receptors add their much larger contribution — so compensation deepens progressively and is maximal at 12–24 hours. The same barrier physiology, running in reverse, is why chronic CO₂ retainers reset their central drive.
- The urine anion gap (U_Na + U_K − U_Cl) is an indirect measure of urinary ammonium, which is not routinely measured. Ammonium is a cation excreted mainly with chloride, so a large ammonium excretion drives urinary chloride up and makes the calculated gap negative. A positive urine anion gap means little ammonium is being excreted — the kidney is failing to generate new bicarbonate — which, combined with an inappropriately high urine pH of 5.9 in the face of systemic acidosis, indicates a renal tubular acidosis (a distal, type 1, picture if the urine cannot be acidified below 5.5). The kidney is the cause of the acidosis. In diarrhea, the kidney is healthy and responding correctly: bicarbonate is being lost through the gut, the kidney maximally increases ammoniagenesis, urinary ammonium is high, the urine anion gap is negative, and the urine pH is appropriately low. Both patients have a normal-anion-gap metabolic acidosis with hyperchloremia and identical basic chemistry; the urine distinguishes a kidney that is the problem from a kidney that is the solution.
31.8 The Four Primary Disorders and How to Diagnose Them
Vocabulary first
Two suffixes, used precisely, prevent most confusion.
- -emia describes the blood pH. Acidemia is pH < 7.35; alkalemia is pH > 7.45. A patient can only be one of these at a time.
- -osis describes a process that, acting alone, would move the pH in a given direction. Acidosis and alkalosis can coexist in the same patient, and their effects can cancel to produce a normal pH.
That last sentence is why a normal pH never means a normal patient. The salicylate case below has a pH of 7.42 and two life-threatening primary disorders.
The five-step method
╔════════════════════════════════════════════════════════════════════════╗
║ FIVE-STEP INTERPRETATION OF ANY ARTERIAL BLOOD GAS ║
╚════════════════════════════════════════════════════════════════════════╝
STEP 0 (free) · CHECK INTERNAL CONSISTENCY
[H+] = 24 × PaCO2 / HCO3- · compare with the reported pH
Disagreement → wrong sample, venous gas, or transcription error.
│
▼
┌──────────────────────────────────────────────────────────────────────┐
│ STEP 1 · LOOK AT THE pH │
│ < 7.35 ACIDEMIA │ 7.35–7.45 NORMAL │ > 7.45 ALKALEMIA │
│ │ (may still hide a │ │
│ │ MIXED disorder — │ │
│ │ go on to step 2) │ │
└──────────┬────────────────────┬─────────────────────┬────────────────┘
▼ ▼ ▼
┌──────────────────────────────────────────────────────────────────────┐
│ STEP 2 · WHICH VALUE EXPLAINS THE pH? = THE PRIMARY DISTURBANCE │
│ │
│ ACIDEMIA + high PaCO2 (>45) ────► RESPIRATORY ACIDOSIS │
│ ACIDEMIA + low HCO3- (<22) ────► METABOLIC ACIDOSIS │
│ ALKALEMIA + low PaCO2 (<35) ────► RESPIRATORY ALKALOSIS │
│ ALKALEMIA + high HCO3- (>26) ───► METABOLIC ALKALOSIS │
│ │
│ ★ If PaCO2 and HCO3- move in OPPOSITE directions from normal │
│ (one up, one down) → there are TWO primary disorders. │
│ ★ COMPENSATION NEVER OVERCORRECTS the pH past 7.40. │
└──────────────────────────────────┬───────────────────────────────────┘
▼
┌──────────────────────────────────────────────────────────────────────┐
│ STEP 3 · IS THE COMPENSATION APPROPRIATE? │
│ │
│ METABOLIC ACIDOSIS expected PaCO2 = 1.5 × HCO3- + 8 ± 2 │
│ (Winter's formula) │
│ METABOLIC ALKALOSIS expected PaCO2 = 40 + 0.7 × (HCO3- − 24) ± 5 │
│ │
│ RESP. ACIDOSIS acute: HCO3- ↑ 1 per 10 mm Hg ↑ PaCO2 │
│ pH ↓ 0.08 per 10 │
│ chronic: HCO3- ↑ 3.5–4 per 10 │
│ pH ↓ 0.03 per 10 │
│ RESP. ALKALOSIS acute: HCO3- ↓ 2 per 10 mm Hg ↓ PaCO2 │
│ pH ↑ 0.08 per 10 │
│ chronic: HCO3- ↓ 4–5 per 10 │
│ pH ↑ 0.03 per 10 │
│ │
│ MEASURED = EXPECTED → simple disorder, appropriately compensated │
│ MEASURED ≠ EXPECTED → a SECOND primary disorder is present │
└──────────────────────────────────┬───────────────────────────────────┘
▼
┌──────────────────────────────────────────────────────────────────────┐
│ STEP 4 · IF THERE IS A METABOLIC ACIDOSIS, CALCULATE THE ANION GAP │
│ │
│ ANION GAP = Na+ − (Cl- + HCO3-) normal 8–12 │
│ ALBUMIN CORRECTION: add 2.5 to the gap for every 1 g/dL that │
│ albumin is below 4.0 g/dL (or lower the "normal" by 2.5) │
│ │
│ HIGH GAP "GOLD MARK" Glycols · Oxoproline (chronic paracetamol) · │
│ L-lactate · D-lactate · Methanol · Aspirin · │
│ Renal failure · Ketoacidosis │
│ NORMAL GAP (hyperchloraemic) diarrhoea · RTA types 1, 2, 4 · │
│ acetazolamide · ureteral diversion · │
│ large-volume 0.9% saline · early CKD │
│ → distinguish with the URINE ANION GAP (UNa + UK − UCl): │
│ NEGATIVE = high urinary NH4+ = GI loss (kidney is fine) │
│ POSITIVE = low urinary NH4+ = RTA (kidney is the problem) │
└──────────────────────────────────┬───────────────────────────────────┘
▼
┌──────────────────────────────────────────────────────────────────────┐
│ STEP 5 · DELTA-DELTA — is a THIRD process hiding? │
│ │
│ Δ ratio = (measured AG − 12) / (24 − measured HCO3-) │
│ │
│ < 0.4 pure NON-gap acidosis │
│ 0.4–0.8 MIXED gap and non-gap acidosis │
│ 1.0–2.0 pure ANION GAP acidosis │
│ > 2.0 coexisting METABOLIC ALKALOSIS, or a pre-existing │
│ chronic respiratory acidosis with a high baseline │
│ │
│ CAVEAT: the ratio runs ~1 in lactic acidosis but often 0.4–0.8 │
│ in DKA, because ketoanions are lost in the urine as Na+/K+ │
│ salts — converting part of the gap acidosis into a non-gap one. │
└──────────────────────────────────────────────────────────────────────┘
Figure 31.9 — A five-step flowchart for interpreting any arterial blood gas, with the compensation formulas and the anion gap and delta-delta rules.
Described: A flowchart with a preliminary step and five numbered steps. Step zero checks internal consistency using the Henderson equation, hydrogen ion concentration equals 24 times the carbon dioxide tension divided by bicarbonate, compared against the reported pH; disagreement indicates a wrong sample, a venous gas, or a transcription error. Step one examines the pH: below 7.35 is acidemia, above 7.45 is alkalemia, and 7.35 to 7.45 is normal but may still conceal a mixed disorder. Step two identifies which value explains the pH and therefore names the primary disturbance: acidemia with a carbon dioxide tension above 45 is respiratory acidosis; acidemia with bicarbonate below 22 is metabolic acidosis; alkalemia with a carbon dioxide tension below 35 is respiratory alkalosis; alkalemia with bicarbonate above 26 is metabolic alkalosis. If carbon dioxide and bicarbonate move in opposite directions from normal, two primary disorders are present, and compensation never overcorrects the pH past 7.40. Step three checks whether compensation is appropriate: for metabolic acidosis, Winter's formula gives expected carbon dioxide tension as 1.5 times bicarbonate plus 8, plus or minus 2; for metabolic alkalosis, expected carbon dioxide tension is 40 plus 0.7 times the rise in bicarbonate above 24, plus or minus 5; for acute respiratory acidosis bicarbonate rises 1 and pH falls 0.08 per 10 mm Hg rise in carbon dioxide, and for chronic respiratory acidosis bicarbonate rises 3.5 to 4 and pH falls 0.03; for acute respiratory alkalosis bicarbonate falls 2 and pH rises 0.08 per 10 mm Hg fall, and for chronic respiratory alkalosis bicarbonate falls 4 to 5 and pH rises 0.03. Agreement between measured and expected indicates a simple, appropriately compensated disorder; disagreement indicates a second primary disorder. Step four, applied when a metabolic acidosis is present, calculates the anion gap as sodium minus the sum of chloride and bicarbonate, normal 8 to 12, corrected upward by 2.5 for every gram per decilitre that albumin falls below 4. High gap causes are given by the mnemonic GOLD MARK — glycols, oxoproline, L-lactate, D-lactate, methanol, aspirin, renal failure, and ketoacidosis. Normal gap hyperchloremic causes include diarrhea, renal tubular acidosis types 1, 2, and 4, acetazolamide, ureteral diversion, large-volume normal saline, and early chronic kidney disease; these are distinguished by the urine anion gap, urinary sodium plus potassium minus chloride, where a negative value indicates high urinary ammonium and gastrointestinal loss with a healthy kidney, and a positive value indicates low urinary ammonium and renal tubular acidosis. Step five calculates the delta ratio, the rise in anion gap above 12 divided by the fall in bicarbonate below 24: under 0.4 indicates a pure non-gap acidosis, 0.4 to 0.8 a mixed gap and non-gap acidosis, 1.0 to 2.0 a pure anion gap acidosis, and above 2.0 a coexisting metabolic alkalosis or a pre-existing chronic respiratory acidosis. A caveat notes the ratio runs near 1 in lactic acidosis but often 0.4 to 0.8 in diabetic ketoacidosis because ketoanions are lost in the urine as sodium and potassium salts.
The four disorders
| Respiratory acidosis | Respiratory alkalosis | Metabolic acidosis | Metabolic alkalosis | |
|---|---|---|---|---|
| Primary change | PaCO₂ ↑ | PaCO₂ ↓ | HCO₃⁻ ↓ | HCO₃⁻ ↑ |
| Underlying event | Hypoventilation | Hyperventilation | Acid gained or HCO₃⁻ lost | H⁺ lost or HCO₃⁻ gained |
| Compensation | Renal: ↑ HCO₃⁻ (days) | Renal: ↓ HCO₃⁻ (days) | Respiratory: ↓ PaCO₂ (minutes–hours) | Respiratory: ↑ PaCO₂ (minutes) |
| Main causes | CNS depression (opioids, sedatives, stroke, OSA); neuromuscular (Guillain-Barré, myasthenia, ALS, high cord injury, severe hypokalemia or hypophosphatemia); chest wall (kyphoscoliosis, obesity hypoventilation, flail chest); airway/lung (COPD, severe asthma, late ARDS) | Anxiety and pain; hypoxemia (PE, pneumonia, altitude, heart failure); sepsis (early); CNS lesions; salicylates; pregnancy (normal PaCO₂ 28–32); liver failure; over-ventilation | High gap: ketoacidosis, lactic acidosis, renal failure, toxic alcohols, salicylate. Normal gap: diarrhea, RTA, acetazolamide, saline | Vomiting or NG suction; diuretics; hypokalemia; mineralocorticoid excess; alkali load; contraction |
| Clinical picture | Somnolence, headache (cerebral vasodilation), asterixis, papilledema, CO₂ narcosis | Perioral and digital paresthesia, carpopedal spasm and tetany (via ionized calcium), light-headedness (cerebral vasoconstriction ~2% per mm Hg fall) | Kussmaul respiration, hypotension below pH 7.1, arrhythmia, hyperkalemia in mineral acidoses, chronic bone demineralization | Often silent; tetany, hypoventilation, arrhythmia, worsening hypokalemia, impaired tissue O₂ unloading |
Metabolic alkalosis needs two things, and this is the part usually left out. A normal kidney can excrete an enormous bicarbonate load within hours. So a sustained metabolic alkalosis requires both a generating cause and a maintaining cause.
| Generation | Maintenance (why the kidney cannot fix it) |
|---|---|
| Loss of H⁺: vomiting, NG suction (every mEq of HCl lost leaves an mEq of HCO₃⁻ behind), diuretics, hyperaldosteronism | Chloride depletion — pendrin in type B intercalated cells cannot secrete HCO₃⁻ without luminal Cl⁻ to exchange it for |
| Gain of alkali: bicarbonate therapy, citrate in massive transfusion, milk-alkali | Volume depletion — avid proximal Na⁺ (and therefore HCO₃⁻) reabsorption; RAAS activation |
| Intracellular shift of H⁺: hypokalemia | Potassium depletion — drives H⁺ into cells and stimulates the H⁺/K⁺-ATPase and ammoniagenesis |
| Contraction: loss of Cl⁻-rich, HCO₃⁻-poor fluid leaves the same HCO₃⁻ in a smaller ECF | Mineralocorticoid excess — continued distal H⁺ and K⁺ secretion |
Classify by urine chloride: < 20 mEq/L = chloride-responsive (vomiting, NG suction, prior diuretic use, post-hypercapnia) and treated with sodium chloride and potassium chloride; > 20 mEq/L = chloride-resistant (primary hyperaldosteronism, Cushing syndrome, Bartter and Gitelman syndromes, severe potassium depletion, current diuretic action) and treated by addressing the cause. Note that urine sodium is unreliable here — bicarbonate being excreted drags sodium with it, so urinary sodium can be high in a volume-depleted patient. Chloride is the honest measurement.
Six worked interpretations
Predict This
Below are two arterial blood gases.
- Patient A: pH 7.42, PaCO₂ 22, HCO₃⁻ 14.
- Patient B: pH 7.26, PaCO₂ 78, HCO₃⁻ 34.
One of them has a completely normal pH and two life-threatening primary acid-base disorders. The other has an obviously abnormal pH and only one. Before reading on, decide which is which, and how you could tell without knowing anything about the patients.
(Answer: Patient A. Both of A's values are far outside their reference ranges and they cancel; no single disorder plus its compensation produces a pH that lands exactly on 7.42 with a bicarbonate of 14 — the compensation rules below prove it. Patient B's pattern is exactly what a single chronic respiratory acidosis with an acute worsening looks like. A normal pH is not reassurance; it is a question.)
Clinical Connection · Six Fully Worked Arterial Blood Gases
WORKED ABG 1 — AMARA (Case File 31). pH 7.49 · PaCO₂ 47 · HCO₃⁻ 34 · Na⁺ 133 · K⁺ 2.9 · Cl⁻ 92 · albumin 3.8
- Step 0 · Consistency. [H⁺] = 24 × 47/34 = 33 nmol/L → pH ≈ 7.48. Reported 7.49. Consistent.
- Step 1 · pH 7.49 → ALKALEMIA.
- Step 2 · Primary disturbance. HCO₃⁻ 34 is high, which would raise pH → fits. PaCO₂ 47 is high, which would lower pH → does not fit, so it must be compensation. Primary metabolic alkalosis.
- Step 3 · Compensation. Expected PaCO₂ = 40 + 0.7 × (34 − 24) = 40 + 7 = 47 (± 5, so 42–52). Measured 47. Exactly appropriate. This is a simple, fully compensated metabolic alkalosis — no second disorder.
- Step 4 · Anion gap. Not strictly required (no metabolic acidosis), but calculate it anyway: 133 − (92 + 34) = 7. Corrected for albumin 3.8: 7 + 2.5 × 0.2 = 7.5. Normal. No hidden gap acidosis.
- Step 5 · Delta-delta. Not applicable.
- Diagnosis: Simple, appropriately compensated, chloride-responsive metabolic alkalosis, with hypokalemia, hypomagnesemia, mild hyponatremia, and prerenal azotemia (BUN:Cr = 32/1.6 = 20), all secondary to loop diuresis on a background of an ACE inhibitor and reduced renal reserve.
- Treatment logic: hold or reduce the furosemide; give potassium chloride (correcting the potassium and the chloride, so pendrin can finally excrete the bicarbonate); give magnesium first or simultaneously, or the potassium will not correct (§31.5); rehydrate cautiously with a chloride-containing fluid; recheck the ECG.
WORKED ABG 2 — DIABETIC KETOACIDOSIS. pH 7.14 · PaCO₂ 22 · HCO₃⁻ 7 · Na⁺ 132 · K⁺ 5.4 · Cl⁻ 95 · glucose 620 · albumin 4.0
- Step 0. [H⁺] = 24 × 22/7 = 75 nmol/L → pH ≈ 7.12. Consistent with 7.14.
- Step 1 · ACIDEMIA, severe.
- Step 2. HCO₃⁻ 7 is low → primary metabolic acidosis.
- Step 3. Winter's: expected PaCO₂ = 1.5 × 7 + 8 = 18.5 (± 2, so 16.5–20.5). Measured 22 — higher than expected. A higher-than-expected PaCO₂ means insufficient respiratory compensation, which is itself a respiratory acidosis superimposed. In practice this means the patient is tiring, or has been given a sedative, and it is an ominous sign.
- Step 4 · Anion gap = 132 − (95 + 7) = 30. Markedly high.
- Step 5 · Delta ratio = (30 − 12)/(24 − 7) = 18/17 = 1.06 → a pure anion gap acidosis; no additional non-gap acidosis or metabolic alkalosis.
- Also: corrected sodium = 132 + 2.4 × (620 − 100)/100 = 144.5 → she is actually hypernatremic and has a large free water deficit hiding behind a "normal-low" sodium.
- Diagnosis: High-anion-gap metabolic acidosis with inadequate respiratory compensation — that is, a superimposed respiratory acidosis. Hypertonic hyponatremia. And, despite a potassium of 5.4, a profound total body potassium deficit (§31.4).
WORKED ABG 3 — COPD WITH ACUTE DECOMPENSATION. pH 7.26 · PaCO₂ 78 · HCO₃⁻ 34 · Na⁺ 140 · Cl⁻ 96
- Step 0. [H⁺] = 24 × 78/34 = 55 nmol/L → pH ≈ 7.26. Consistent.
- Step 1 · ACIDEMIA.
- Step 2. PaCO₂ 78 is high → fits an acidemia. HCO₃⁻ 34 is high → would raise pH, so it is compensation. Primary respiratory acidosis.
- Step 3 · Acute or chronic? ΔPaCO₂ = 78 − 40 = 38, i.e. 3.8 "tens."
- If acute: expected HCO₃⁻ = 24 + 1 × 3.8 = 27.8, expected pH = 7.40 − 0.08 × 3.8 = 7.10.
- If chronic: expected HCO₃⁻ = 24 + 3.5 × 3.8 = 37.3, expected pH = 7.40 − 0.03 × 3.8 = 7.29.
- Measured HCO₃⁻ 34 and pH 7.26 both sit between the two predictions, closer to the chronic. Acute-on-chronic respiratory acidosis.
- Step 4 · Anion gap = 140 − (96 + 34) = 10. Normal — no metabolic acidosis hiding.
- Diagnosis: Acute-on-chronic respiratory acidosis. The renal compensation proves the CO₂ retention has been present for days at least; the pH proves it has recently worsened.
- Treatment logic: treat the acute precipitant; non-invasive ventilation is aimed at the pH, not the PaCO₂ — his baseline PaCO₂ may be 60 and driving it to 40 would produce a post-hypercapnic metabolic alkalosis. Target SpO₂ 88–92%.
WORKED ABG 4 — SALICYLATE OVERDOSE. pH 7.42 · PaCO₂ 22 · HCO₃⁻ 14 · Na⁺ 140 · Cl⁻ 102 · albumin 4.0
- Step 0. [H⁺] = 24 × 22/14 = 38 nmol/L → pH ≈ 7.42. Consistent.
- Step 1 · pH 7.42 — NORMAL. Do not stop here.
- Step 2. Both values are far from normal, and they oppose each other: a PaCO₂ of 22 alone would give a pH near 7.6, and a HCO₃⁻ of 14 alone would give a pH near 7.2. Two primary disorders.
- Step 3 · Test each possibility.
- If the metabolic acidosis were primary: Winter's expected PaCO₂ = 1.5 × 14 + 8 = 29 (± 2). Measured 22 — far below. → an additional primary respiratory alkalosis.
- If a chronic respiratory alkalosis were primary: expected HCO₃⁻ = 24 − 4.5 × 1.8 = 16. Measured 14 — below. → an additional metabolic acidosis.
- Step 4 · Anion gap = 140 − (102 + 14) = 24. High, confirming a genuine anion gap metabolic acidosis rather than compensation.
- Step 5 · Delta ratio = (24 − 12)/(24 − 14) = 12/10 = 1.2 → the acidosis is a pure gap acidosis.
- Diagnosis: Mixed primary respiratory alkalosis and primary high-anion-gap metabolic acidosis — the classic salicylate pattern (§31.9). The normal pH is the most dangerous feature of this gas, because it invites reassurance.
WORKED ABG 5 — PROLONGED VOMITING (pyloric obstruction). pH 7.52 · PaCO₂ 45 · HCO₃⁻ 36 · Na⁺ 138 · K⁺ 3.0 · Cl⁻ 88 · urine Cl⁻ 8 mEq/L
- Step 0. [H⁺] = 24 × 45/36 = 30 nmol/L → pH ≈ 7.52. Consistent.
- Step 1 · ALKALEMIA.
- Step 2. HCO₃⁻ 36 high → primary metabolic alkalosis.
- Step 3. Expected PaCO₂ = 40 + 0.7 × 12 = 48.4 (± 5, so 43–53). Measured 45. Appropriate. Simple disorder.
- Step 4 · Anion gap = 138 − (88 + 36) = 14 — mildly raised. Two ordinary explanations: alkalemia increases the net negative charge on albumin (each 0.1 pH unit adds roughly 1–3 to the gap), and volume depletion generates a little lactate. Not a separate disorder.
- Urine chloride 8 → chloride-responsive. Treatment: sodium chloride and potassium chloride. Note the "paradoxical aciduria" that develops in these patients (§31.9).
- Diagnosis: Simple, appropriately compensated, chloride-responsive metabolic alkalosis with hypokalemia from renal potassium wasting.
WORKED ABG 6 — SEVERE DIARRHEA IN A CHILD. pH 7.28 · PaCO₂ 28 · HCO₃⁻ 13 · Na⁺ 136 · K⁺ 3.1 · Cl⁻ 113 · albumin 4.0
- Step 0. [H⁺] = 24 × 28/13 = 52 nmol/L → pH ≈ 7.29. Consistent.
- Step 1 · ACIDEMIA.
- Step 2. HCO₃⁻ 13 low → primary metabolic acidosis.
- Step 3. Winter's: expected PaCO₂ = 1.5 × 13 + 8 = 27.5 (± 2). Measured 28. Appropriate — respiratory compensation is working perfectly.
- Step 4 · Anion gap = 136 − (113 + 13) = 10. NORMAL → a hyperchloremic (non-gap) metabolic acidosis. Note the chloride of 113: as bicarbonate is lost, chloride is retained to preserve electroneutrality.
- Step 5 · Delta ratio = (10 − 12)/(24 − 13) = −0.18, i.e. < 0.4 → confirms a pure non-gap acidosis.
- Urine anion gap would be negative, confirming high urinary ammonium and therefore that the kidney is working correctly and the bicarbonate is being lost through the gut.
- Diagnosis: Normal-anion-gap hyperchloremic metabolic acidosis with appropriate respiratory compensation and hypokalemia, from loss of bicarbonate- and potassium-rich intestinal fluid.
BONUS — WORKED ABG 7 — POST-CARDIAC-ARREST. pH 6.99 · PaCO₂ 60 · HCO₃⁻ 14 · Na⁺ 138 · Cl⁻ 100 · lactate 12 mmol/L
- Step 1 · Profound ACIDEMIA.
- Step 2. PaCO₂ is high and HCO₃⁻ is low — both abnormalities push the pH in the same direction. There is no compensation at all; there are two primary acidoses.
- Step 3. Winter's for HCO₃⁻ 14 predicts a PaCO₂ of 29; it is 60. Acute respiratory acidosis at a PaCO₂ of 60 predicts a HCO₃⁻ of 26; it is 14. Each confirms the other.
- Step 4 · Anion gap = 138 − (100 + 14) = 24, from the lactate.
- Step 5 · Delta ratio = 12/10 = 1.2 → pure gap acidosis.
- Diagnosis: Combined high-anion-gap metabolic acidosis (lactic, from arrest-related tissue hypoperfusion) and respiratory acidosis (from inadequate ventilation during and after the arrest). Treatment is restoring perfusion and ventilation; bicarbonate does not treat either and can worsen intracellular acidosis by generating CO₂ that diffuses into cells.
31.9 Advanced Topic · Integrated Disturbances
Each of the following is a complete multi-system problem in which fluid, electrolyte, and acid-base derangements arise from a single primary event and then interact.
Diabetic ketoacidosis
The chain. Absolute or relative insulin deficiency, with excess glucagon, catecholamines, cortisol, and growth hormone → unrestrained lipolysis → free fatty acids flood the liver → β-oxidation generates acetyl-CoA far faster than the citric acid cycle can consume it (and oxaloacetate is being diverted to gluconeogenesis) → acetyl-CoA is condensed into acetoacetate and β-hydroxybutyrate → a high-anion-gap metabolic acidosis. Simultaneously, absent insulin plus gluconeogenesis raise glucose above the renal threshold of ~180 mg/dL → an osmotic diuresis.
The losses, which are enormous and routinely underestimated:
| Typical total deficit | |
|---|---|
| Water | 5–8 L (~100 mL/kg) |
| Sodium | 400–700 mEq (7–10 mEq/kg) |
| Potassium | 300–1,000 mEq (3–15 mEq/kg) |
| Phosphate | 50–100 mmol (~1 mmol/kg) |
| Magnesium | 25–50 mEq |
The potassium trap, in full. Total body potassium is profoundly depleted — hours of osmotic diuresis with high distal flow and ketoanion delivery (levers 3 and 4 of Figure 31.4C), plus vomiting, plus secondary hyperaldosteronism from volume depletion. Yet serum potassium is often normal or high, because insulin deficiency, hypertonicity, and (to a lesser degree) the acidosis have all shifted potassium out of cells. Give insulin and the shift reverses within minutes; serum potassium can drop 1–2 mEq/L in the first hour and precipitate arrhythmia.
Hence the treatment order: fluid, then potassium, then insulin. - Fluid first, because volume restoration alone lowers glucose substantially (by restoring GFR and hence glucose excretion) and improves tissue perfusion. - Potassium before insulin if K⁺ < 3.3 mEq/L. Between 3.3 and 5.3, give potassium with the insulin. Above 5.3, hold potassium and recheck hourly. - Insulin by continuous infusion, which does two things: it switches off ketogenesis (the actual disease) and it drives glucose into cells (the visible number).
Two endpoints that are commonly confused. The glucose normalizes long before the ketosis resolves — which is why dextrose is added to the fluids once glucose falls below ~200 mg/dL while the insulin infusion continues. The endpoint is closure of the anion gap, not the glucose. And in late treatment a hyperchloremic non-gap acidosis typically appears: large volumes of chloride-rich saline have been given, and the ketoanions that would have regenerated bicarbonate on metabolism were excreted in the urine as sodium and potassium salts instead. It is expected, benign, and resolves over a day or two.
Vomiting versus diarrhea — the same organ system, opposite results
This pair is the best single illustration in the chapter of why which fluid is lost matters more than how much.
| Vomiting / nasogastric suction | Diarrhea | |
|---|---|---|
| Fluid lost | Gastric: H⁺ ~60–100 mEq/L, Cl⁻ ~100–140, K⁺ 5–10 | Intestinal/pancreatic: HCO₃⁻ 30–70 mEq/L, K⁺ 30–90 in colonic fluid |
| Acid-base result | Metabolic ALKALOSIS | Metabolic ACIDOSIS, normal gap |
| Chloride | Depleted → alkalosis maintained | Normal or high (hyperchloremic) |
| Potassium | Low, mostly by renal loss | Low, by direct loss |
| Anion gap | Normal (mildly raised by alkalemia) | Normal |
| Urine anion gap | — | Negative (kidney making ammonium correctly) |
The vomiting sequence in detail. Every mEq of HCl secreted into the stomach leaves an mEq of HCO₃⁻ behind in the blood — the "alkaline tide," normally cancelled when the acid is neutralized by pancreatic bicarbonate in the duodenum. Vomiting removes the acid before that cancellation, so the bicarbonate stands unopposed: the alkalosis is generated. Then volume and chloride depletion maintain it: the kidney avidly reabsorbs sodium (and bicarbonate with it), pendrin cannot excrete bicarbonate without luminal chloride, and aldosterone drives distal H⁺ and K⁺ secretion.
That last item produces the finding that confuses everyone: paradoxical aciduria. A patient with an arterial pH of 7.55 producing acid urine. It looks like a failure of regulation and it is not — it is a correctly prioritized conflict. The kidney has been asked to defend volume and potassium and pH simultaneously, and volume wins. Understanding why requires exactly the volume-versus-concentration distinction of §31.3.
COPD with acute decompensation
The chronic state: alveolar hypoventilation from airflow obstruction raises PaCO₂ over months to years. The kidney compensates fully — inducing ammoniagenesis over days to weeks — raising bicarbonate to 32–38 and returning pH close to normal. The patient lives, quite safely, at a PaCO₂ of 55–65.
Then something acute happens: a chest infection, a sedative, heart failure, or a fractured rib. PaCO₂ rises further within hours, and now there is no time for renal compensation — which takes days. Bicarbonate is already near its maximum. The pH falls steeply.
Why oxygen can worsen it — a mechanism worth getting right, because the traditional teaching is largely wrong:
- Reversal of hypoxic pulmonary vasoconstriction (the dominant mechanism). Poorly ventilated lung units are normally vasoconstricted, diverting perfusion to better-ventilated regions. High inspired oxygen relieves that constriction, perfusion returns to poorly ventilated units, and physiological dead space increases — so the same minute ventilation clears less CO₂.
- The Haldane effect. Oxygenated hemoglobin binds CO₂ and H⁺ less avidly than deoxygenated hemoglobin, so oxygenating the blood displaces CO₂ into the plasma, raising PaCO₂ at any given ventilation.
- Reduced hypoxic ventilatory drive — real, but a minority contributor.
Hence the target of SpO₂ 88–92% in known CO₂ retainers: enough oxygen to prevent tissue hypoxia, not so much that dead space and the Haldane effect drive PaCO₂ up. And hypoxia, not hypercapnia, is what kills quickly — so oxygen is never withheld from a hypoxic patient. It is titrated.
Salicylate overdose — the classic mixed disorder
Aspirin does three separate things, on three different timescales, and they produce three different acid-base disorders in the same patient.
- Direct stimulation of the medullary respiratory center → hyperventilation → primary respiratory alkalosis, appearing within the first hours. This is the earliest finding, and in adults it may dominate the picture.
- Uncoupling of oxidative phosphorylation and inhibition of Krebs cycle enzymes → mitochondria burn substrate without making ATP → heat production rises (hyperthermia), anaerobic glycolysis rises (lactate), lipolysis rises (ketones), plus the salicylate anion itself → primary high-anion-gap metabolic acidosis.
- Late respiratory acidosis from respiratory muscle fatigue, cerebral edema, or — most often — iatrogenic sedation and intubation.
That third item is the lethal turn, and it depends on a piece of chemistry. Salicylic acid has a pKa of about 3.0, so at pH 7.4 it is over 99.9% ionized and cannot cross membranes. As pH falls, the un-ionized fraction rises exponentially, and un-ionized salicylate crosses into the central nervous system. So a falling pH does not merely accompany deterioration — it causes it, by moving the poison into the brain.
The treatment logic follows exactly:
- Alkalinize. Sodium bicarbonate raises serum pH, keeping salicylate ionized and out of the brain, and raises urine pH above 7.5, trapping ionized salicylate in the tubular lumen for excretion — increasing clearance several-fold.
- Replace potassium, because you cannot alkalinize the urine in a hypokalemic patient: the kidney will reabsorb potassium in exchange for secreting H⁺, acidifying the urine no matter how much bicarbonate you give. This is §31.4's α-intercalated cell H⁺/K⁺-ATPase, deciding the outcome of a poisoning.
- Give glucose regardless of serum glucose, because CNS glucose utilization can be impaired with a normal blood level.
- Do not intubate casually. Paralysis abolishes the compensatory hyperventilation, PaCO₂ rises immediately toward 40, pH crashes, salicylate floods into the brain, and patients have arrested at this moment. If intubation is unavoidable, hyperventilate deliberately and give bicarbonate through the transition.
- Hemodialysis for severe toxicity, altered mental status, renal failure, or pulmonary edema.
Exercise-associated hyponatremia
Exercise & Sport · Exercise-Associated Hyponatremia and "Drink to Thirst"
For most of the late twentieth century athletes were told to drink aggressively and stay "ahead of thirst." That advice caused deaths, and the physiology explaining why is a clean application of §31.2 and §31.3.
The mechanism has two halves, and the first is the bigger one.
- Excess intake of hypotonic fluid. Water or sports drink (typical sodium 10–20 mEq/L, far below plasma) is consumed faster than it can be excreted. The clearest marker is that the athlete gains weight during the event — an unmistakable sign of positive fluid balance in someone who should be losing 1–3 kg.
- Non-osmotic ADH release. Exercise itself, plus pain, nausea, hypoglycemia, and NSAIDs, all drive ADH secretion regardless of osmolality — the baroreceptor and stress override of Figure 31.3. So at exactly the moment the athlete needs to excrete free water, her kidney is being told to retain it. Maximum free water clearance, normally 800–1,000 mL/h, may fall to a fraction of that.
Sodium losses in sweat contribute but are usually secondary, because sweat is hypotonic (§31.4) and sweating alone therefore raises plasma sodium.
The numbers. Depending on the event and the definition, up to 13% of marathon finishers have been reported with a sodium below 135 mEq/L; the great majority are asymptomatic. Symptomatic cases are rare but potentially fatal, from cerebral edema and — characteristically — a non-cardiogenic pulmonary edema that accompanies it. Risk factors: event duration over four hours, low body mass, slower finishing times (more drinking opportunities), female sex, NSAID use, and unrestricted fluid availability.
Why the treatment is counter-intuitive. A collapsed athlete with hyponatremia looks dehydrated, and the reflex is to give intravenous saline. In this setting that can make things worse. With ADH high, the kidney retains the water from isotonic saline while excreting the sodium in a concentrated urine — "desalination" — so the net effect of a litre of 0.9% saline can be to lower the serum sodium further. The correct treatment for symptomatic exercise-associated hyponatremia is hypertonic saline: 100 mL of 3% NaCl, repeated up to three times, aiming to raise sodium by 4–5 mEq/L, which is enough to reverse cerebral edema. Note that the usual slow-correction rule does not apply, because this hyponatremia is acute — the brain has had no time to extrude osmolytes, so there is no risk of osmotic demyelination and every reason to correct quickly.
Current guidance: drink to thirst. The osmoreceptor–ADH–thirst loop is a well-calibrated, fast, and accurate controller, and in a healthy adult with access to fluid it substantially outperforms any schedule. Thirst begins at an osmolality of about 290–295 mOsm/kg, which corresponds to roughly 2% dehydration — a level associated with essentially no performance decrement. Drinking ahead of thirst overrides a working control system on the basis of a worse estimate.
And the counterexample that keeps it honest: in heat, with sweat rates of 1.5–2.5 L/h, thirst may under-replace over many hours — the "voluntary dehydration" that is well documented in laboratory heat studies. So the guidance is drink to thirst and monitor body weight over multi-day heat exposure, which is exactly what military and occupational heat protocols do.
Heat illness completes the picture. Heat exhaustion is a compensating state: sweating continues, core temperature is under 40 °C, and the problem is volume and salt depletion with cardiovascular strain — treated with rest, cooling, and oral fluid with salt. Heat stroke is decompensated: core temperature above 40 °C with central nervous system dysfunction, often with hot dry skin because sweating has failed, and it is a true emergency requiring immediate cooling — cold water immersion lowers core temperature fastest, and time above 40 °C is what determines organ damage. The distinction is the one Chapter 1 drew for thermoregulation: is the homeostatic loop still working, or has it been overwhelmed?
Thread 3 · The Body Is Integrated
Look back at Amara's single sheet of numbers. A diuretic given for a cardiovascular problem (heart failure after an infarction) acted on a renal transporter, which changed electrolyte handling at three different nephron segments, which produced an acid-base disorder, which altered protein binding of calcium, which — together with the potassium and magnesium — changed neuromuscular and cardiac excitability, which produced her symptoms: weakness, cramps, and palpitations. The hepatic metabolism of lactate and the pulmonary control of CO₂ are both participating in the compensation as you read this.
Seven systems, one prescription, one blood tube.
This is the chapter where the book's threads converge, and the convergence is quantitative. You cannot reason your way through Amara's sheet with any single chapter of this book. You can reason your way through it with all of them, and the reasoning is arithmetic that you can now do.
Chapter Summary
§31.1 Total body water is 50–60% of mass in adults, 75–80% in newborns, and 45–50% after 75 — the variation tracks adiposity and muscle. Two-thirds is intracellular (28 L in a 70 kg adult), one-third extracellular (14 L: 10.5 L interstitial, 3.5 L plasma). All compartments share one osmolality (~285 mOsm/kg) and differ completely in composition, because the Na⁺/K⁺ ATPase makes sodium extracellular and potassium intracellular at 20–30% of resting metabolic cost. Tonicity, not osmolality, moves water: urea is an ineffective osmole; glucose is not. One litre of isotonic saline stays entirely in the ECF and adds ~250 mL to plasma; one litre of D5W is free water and adds ~83 mL; one litre of 3% saline expands the ECF by 2.6 L by drawing 1.6 L out of cells.
§31.2 Intake ~2,500 mL/day including ~300 mL of metabolic water; obligatory loss ~1,400 mL, of which the urinary floor is set by the daily solute load divided by maximum urine concentration (600 ÷ 1.2 = 500 mL in youth; 600 ÷ 0.7 = 857 mL at 80). The osmoreceptor–ADH–thirst loop defends osmolality to within 1–2%: osmoreceptors in the circumventricular organs, an ADH threshold near 280–285 and a thirst threshold near 290–295, with ADH inserting aquaporin-2 into collecting duct principal cells. A volume stimulus overrides the osmotic one, which is the origin of most clinical hyponatremia. Edema is a Starling-force problem with four possible causes and requires 2.5–3 L of excess interstitial fluid before it is visible.
§31.3 Sodium determines ECF volume; water determines osmolality and therefore serum sodium concentration. The serum sodium tells you nothing about total body sodium. Sodium is controlled by RAAS, ANP/BNP, sympathetic tone, and pressure natriuresis; water by ADH and thirst. Hyponatremia is classified first by tonicity, then by volume status, and hypervolemic hyponatremia means too much sodium and even more water. Correcting hyponatremia faster than ~8 mEq/L per 24 h risks osmotic demyelination; correcting hypernatremia faster than ~10–12 mEq/L per 24 h risks cerebral edema. In both, the injury is caused by the correction.
§31.4 98% of potassium is intracellular; the ECF holds ~70 mEq and plasma only ~15–20, so small shifts are large concentration changes. The 35 : 1 gradient sets E_K at −94 mV. Hyperkalemia depolarizes and inactivates sodium channels; hypokalemia hyperpolarizes but slows repolarization because I_K1 conductance is proportional to √[K⁺]ₒ — hence U waves and torsades. Shifts (insulin, β₂-agonists, pH, tonicity, lysis, exercise) change the concentration; intake and renal excretion change the total. Collecting duct secretion has four levers: aldosterone, distal sodium delivery, tubular flow, and non-reabsorbable luminal anion. A loop diuretic pulls all four.
§31.5 Ionized calcium is the regulated fraction; alkalosis lowers it at an unchanged total calcium because H⁺ and Ca²⁺ compete for albumin binding sites, producing tetany with a normal laboratory calcium. PTH raises calcium and lowers phosphate; calcitriol raises both; FGF23 lowers phosphate. Magnesium is required for Mg-ATP (and hence the Na⁺/K⁺ ATPase), for PTH secretion and action, and to block ROMK from the cytoplasmic side — so magnesium depletion causes unrestrained renal potassium wasting and makes both hypokalemia and hypocalcemia refractory until it is corrected. Refeeding drives phosphate, potassium, and magnesium into cells within 24–72 h.
§31.6 Normal [H⁺] is 40 nmol/L — 3.5 million times lower than sodium — and pH matters because protein charge, and therefore protein shape, depends on it. The body handles ~15,000 mmol/day of volatile acid (CO₂, exhaled) and ~70 mEq/day of fixed acid (excreted renally). Three lines of defense: buffers in seconds (bicarbonate, hemoglobin, protein, phosphate, bone), respiratory compensation in minutes to hours, renal compensation in hours to days. pH depends on the ratio of bicarbonate to dissolved CO₂ — normally 20 : 1 — and the bicarbonate system works despite a pKa of 6.1 because it is an open system with two independently regulated members.
§31.7 Ventilation and PaCO₂ are inversely proportional; central chemoreceptors sense CSF [H⁺] (CO₂ crosses the barrier, H⁺ and HCO₃⁻ do not), which is why respiratory compensation for a metabolic disturbance takes 12–24 h to peak. The kidney does three separate jobs: reclaims ~4,300 mEq/day of filtered bicarbonate (generating none), generates new bicarbonate by excreting titratable acid on phosphate (10–40 mEq/day), and generates new bicarbonate by excreting ammonium (30–40 mEq/day, inducible to 300–400 over 3–5 days). Type A intercalated cells secrete acid; type B are their mirror image and secrete base through pendrin, which requires luminal chloride — the molecular basis of chloride-responsive alkalosis.
§31.8 Five steps: pH; primary disturbance; expected compensation (Winter's formula and the respiratory rules); anion gap if there is a metabolic acidosis; delta-delta for a third process. Compensation never overcorrects, and a normal pH does not exclude two severe primary disorders. Metabolic alkalosis requires both a generating and a maintaining cause, and urine chloride separates chloride-responsive from chloride-resistant forms. Seven fully worked examples apply the method.
§31.9 Integrated disturbances: DKA (high-gap acidosis with massive total body potassium depletion despite a normal or high serum value — fluid, then potassium, then insulin, with the anion gap as the endpoint); vomiting versus diarrhea (the same organ system producing opposite acid-base results, with paradoxical aciduria explained by prioritized conflict); COPD decompensation (why oxygen raises CO₂, mostly through dead space and the Haldane effect); salicylate (three disorders on three timescales, with pH determining CNS penetration); and exercise-associated hyponatremia (overdrinking plus non-osmotic ADH; treated with hypertonic saline, prevented by drinking to thirst).
The Three Threads in Chapter 31
Structure → Function. The whole chapter rests on structural facts. A membrane permeable to water but not to sodium makes tonicity a physical force. A capillary wall permeable to small solutes but not to protein makes plasma a separate compartment. A collecting duct principal cell with ENaC on one face and ROMK on the same face makes potassium secretion depend on sodium delivery. A tubular epithelium whose type A and type B intercalated cells are mirror images makes the kidney able to excrete acid or base with the same machinery. Look at the transporter map and the physiology is already written.
Homeostasis. This is the chapter where the master concept becomes numerical. Every variable here — volume, osmolality, potassium, calcium, pH — is defended by a loop with an identifiable receptor, control center, and effector, and every disorder in the chapter is that loop failing, being overridden, or being correctly prioritized against another loop. Paradoxical aciduria, hypovolemic hyponatremia, and the exercise pressor reflex are all cases of two loops in conflict and the more urgent one winning. That is not a failure of homeostasis. It is what homeostasis looks like when the variables compete.
Integration. One prescription — furosemide — reached seven systems in Amara. And Adwoa's delirium in Chapter 30, which looked like a neurological problem, turns out on this chapter's arithmetic to be a 1.4-litre water deficit in a woman whose kidney could not concentrate her urine and whose hypothalamus did not tell her she was thirsty. The unifying claim of this book is that these are the same subject, and this chapter is where you can finally prove it with numbers.
Case File 31 · Resolution
Question 1 — What acid-base disturbance does Amara have, and is it compensated?
Work the five steps.
Step 0 · Is the gas internally consistent? Using the Henderson equation, [H⁺] = 24 × PaCO₂ / [HCO₃⁻] = 24 × 47 / 34 = 33 nmol/L, which corresponds to a pH of about 7.48 — within measurement error of the reported 7.49. The three values agree, so this is a real arterial sample and worth interpreting.
Step 1 · The pH is 7.49 → ALKALEMIA.
Step 2 · Which value explains it? Her bicarbonate is 34 mEq/L, well above the 22–26 range, and a high bicarbonate raises pH — that fits. Her PaCO₂ is 47 mm Hg, above the 35–45 range, and a high PaCO₂ lowers pH — that does not fit, so it cannot be the primary problem. It must be compensation. The primary disturbance is a metabolic alkalosis.
Step 3 · Is the compensation appropriate? For a metabolic alkalosis, expected PaCO₂ = 40 + 0.7 × (HCO₃⁻ − 24) = 40 + 0.7 × 10 = 47 mm Hg, with an acceptable range of roughly 42–52. Her measured PaCO₂ is exactly 47. The respiratory compensation is precisely what it should be — she is hypoventilating just enough to retain CO₂ and blunt the alkalemia, and no more. This is a simple, appropriately compensated metabolic alkalosis. There is no second primary disorder.
Two things follow from that conclusion that are worth stating explicitly. First, her PaCO₂ of 47 should not be reported as a "respiratory acidosis" — compensation is not a disorder, and calling it one leads people to treat it. Second, because the compensation is appropriate, you do not need to look for a hidden respiratory problem; if her PaCO₂ had been 40 (no compensation) or 58 (too much), you would.
Step 4 · Anion gap. Not required, since there is no metabolic acidosis — but calculate it, as a matter of routine, to make sure nothing is hiding. AG = Na⁺ − (Cl⁻ + HCO₃⁻) = 133 − (92 + 34) = 7. Corrected for her albumin of 3.8: 7 + 2.5 × (4.0 − 3.8) = 7.5. Normal. There is no occult anion gap acidosis being masked by the alkalosis, which is a real and easily missed pattern in sick patients.
Step 5 · Delta-delta. Not applicable in the absence of a metabolic acidosis.
The complete diagnosis, in one sentence: a simple, appropriately compensated, chloride-responsive metabolic alkalosis with hypokalemia, hypomagnesemia, mild hypotonic hyponatremia, and prerenal azotemia (BUN:creatinine = 32/1.6 = 20), all produced by loop diuresis in a patient on an ACE inhibitor with reduced renal reserve.
Where did the alkalosis come from? Four generating mechanisms, all from the furosemide: contraction (loss of chloride-rich, bicarbonate-poor fluid concentrates the remaining bicarbonate in a smaller ECF); secondary hyperaldosteronism from volume depletion, driving distal H⁺ secretion by α-intercalated cells; hypokalemia, which shifts H⁺ into cells and stimulates both the H⁺/K⁺-ATPase and renal ammoniagenesis; and increased distal delivery of sodium with a non-reabsorbable anion, which enhances H⁺ secretion.
And why has it persisted? Because generation is not enough — a normal kidney dumps excess bicarbonate within hours. It persists because she is chloride-depleted (Cl⁻ 92) and potassium-depleted, and type B intercalated cells cannot secrete bicarbonate through pendrin without luminal chloride to exchange it for. She will stay alkalotic until she is given chloride, and the right way to give it is as potassium chloride, which repairs the maintenance mechanism and the potassium deficit with the same molecule. Add magnesium, and the treatment list is complete.
Question 2 — How does a drug acting on the loop of Henle cause a potassium of 2.9?
The premise of the question is correct: the loop is not where potassium is regulated. About 65% of filtered potassium is reabsorbed in the proximal tubule and 25% in the thick ascending limb, but nearly all of what appears in the urine is secreted by the collecting duct, and that is where regulation lives. So the answer is that furosemide's potassium effect is almost entirely downstream of, and secondary to, its site of action — and one mechanism is not in the tubule at all.
Mechanism 1 · The direct loss, which is smaller than it looks. Furosemide inhibits NKCC2 in the thick ascending limb, so potassium that would have been reabsorbed there is not. But most of the potassium NKCC2 brings into the cell is normally recycled straight back into the lumen through apical ROMK, because that recycling is what generates the segment's lumen-positive potential. Blocking the transporter removes both arms. Net direct loss: real, but a minor part of the total.
Mechanism 2 · Increased distal sodium delivery — the dominant mechanism. Sodium not reabsorbed upstream arrives at the principal cells of the late distal tubule and collecting duct. It enters through ENaC. Every sodium ion that leaves the lumen makes the lumen more electrically negative, and it is that negativity which provides the electrical driving force pulling potassium out of the cell through ROMK. More sodium delivered → more sodium through ENaC → more negative lumen → more potassium secreted.
Mechanism 3 · Increased tubular flow. The same blocked reabsorption delivers extra water. High flow sweeps secreted potassium downstream, so luminal potassium stays low and the chemical gradient for further secretion stays maximal; and flow bends the principal cell's primary cilium, raising intracellular calcium and opening BK (maxi-K) channels that are silent at rest. Mechanisms 2 and 3 multiply rather than add, which is why loop diuretics waste potassium so efficiently.
Mechanism 4 · RAAS activation — the mechanism that is not in the tubule. Furosemide depletes ECF volume. That is sensed by afferent arteriolar stretch receptors, by the macula densa (which is doubly affected, since the macula densa's own NKCC2 is inhibited by the drug and therefore reports low NaCl), and by arterial baroreceptors driving renal sympathetic outflow. Renin rises → angiotensin II → aldosterone, which inserts more ENaC, more Na⁺/K⁺ ATPase, and more ROMK into principal cells. This is a hormonal, whole-body loop, and it is why the potassium loss continues long after any single dose has worn off.
Her lisinopril should oppose this — and it does, partially. But the volume depletion stimulus is powerful, aldosterone breakthrough during chronic ACE inhibition is well described, and the net effect is that the furosemide is winning. The lisinopril is also contributing to her creatinine of 1.6: by dilating the efferent arteriole in a volume-depleted patient it lowers glomerular filtration pressure, a hemodynamic effect rather than tubular injury, and one that reverses when volume is restored.
Mechanism 5 · The alkalosis feeds the hypokalemia, and the hypokalemia feeds the alkalosis. Alkalemia shifts potassium into cells, lowering the measured value further. Increased distal bicarbonate delivery acts as a non-reabsorbable luminal anion, deepening lumen negativity and driving still more potassium secretion. In the other direction, potassium depletion shifts H⁺ into cells, upregulates the α-intercalated cell H⁺/K⁺-ATPase (reclaiming potassium at the price of secreting acid), and stimulates ammoniagenesis — all of which sustain the alkalosis. The two abnormalities are mutually reinforcing, and treating either alone fails.
Mechanism 6 · And the one that sets up question 3. The lumen-positive potential of the thick ascending limb also drives paracellular reabsorption of magnesium and calcium through the claudin-16/19 pore. Abolishing that potential wastes magnesium obligatorily. Her magnesium of 1.4 mg/dL is not a coincidence; it is the same drug acting on the same transporter.
Question 3 — Why does potassium replacement fail until magnesium is corrected?
Three reasons, of which the first is decisive.
1 · Intracellular magnesium is the physiological blocker of ROMK. The renal outer medullary potassium channel is an inward rectifier: it conducts potassium into the cell much more readily than out of it, and the reason is that intracellular Mg²⁺ (and polyamines) plug the channel pore from the cytoplasmic side at depolarized potentials, limiting outward — that is, secretory — flux. When intracellular magnesium falls, that block is relieved. ROMK conducts potassium into the tubular lumen without restraint, and renal potassium wasting continues no matter how much potassium you administer. The replacement is filtered, delivered, and secreted straight back out.
2 · The Na⁺/K⁺ ATPase requires Mg-ATP. ATP is functionally never free ATP inside a cell; it is chelated to magnesium, and the pump's substrate is Mg-ATP. Magnesium depletion therefore impairs the pump directly, so cells take up administered potassium less efficiently — the 98% compartment cannot be refilled at the normal rate.
3 · Hypomagnesemia mildly stimulates aldosterone secretion, adding lever 1 of Figure 31.4C to the list.
Therefore: replace magnesium first, or at the same time. And note two practical points. First, a normal serum magnesium does not exclude depletion — less than 1% of body magnesium is extracellular, so serum magnesium is a poor index of total body content, and in a patient with an unexplained refractory hypokalemia on a loop diuretic it is reasonable to replace magnesium even if the level looks acceptable. Second, the same mechanism produces refractory hypocalcemia: magnesium depletion impairs both PTH secretion and the target tissue's response to PTH, so calcium will not correct either until magnesium does. A patient with unexplained, treatment- resistant hypokalemia and hypocalcemia has hypomagnesemia until proven otherwise.
Amara's complete treatment plan, assembled from all three answers: reduce or hold the furosemide; replace magnesium (intravenously, given her level and symptoms); replace potassium as potassium chloride, which simultaneously supplies the chloride her kidney needs to excrete the excess bicarbonate; rehydrate with a chloride-containing isotonic fluid, cautiously given her heart failure; review the lisinopril dose against her creatinine; consider adding a potassium-sparing agent such as spironolactone, which would address the secondary hyperaldosteronism at its source and has independent benefit in heart failure; and repeat the ECG once the potassium is above 3.5.
Systems Integration Case File · Entry 31
Entry 31 — The chemistry underneath everything
Twenty-six chapters of your file describe organs. This entry describes the solution they all sit in, and it is the entry that will make the rest cohere.
New findings. Amara's electrolytes, arterial blood gas, ECG, and drug list, as given in the Case File opener; and, from Chapter 30, Adwoa's sodium of 148 with a BUN of 38 and a creatinine of 1.4 from a baseline of 0.9.
Your entry:
1 · ADD. In no more than four sentences, state Amara's complete acid-base and electrolyte diagnosis, with the reasoning that gets you there — not just the conclusion.
2 · CONNECT. Trace one causal chain from furosemide to palpitations, naming at least four organ systems in order, and stating the direction of causation at every step. Then trace a second, independent chain from furosemide to her weakness, and identify the point where the two chains diverge.
3 · PREDICT. Amara's eGFR is now roughly 40 mL/min/1.73 m². Predict what will happen to her potassium over the next two years as her chronic kidney disease progresses, given that she is on both a loop diuretic and an ACE inhibitor, and say which of the two effects will eventually dominate and why. Then predict one thing you expect in Chapter 33.
Model responses — read only after writing your own
1 · ADD. A pH of 7.49 with a bicarbonate of 34 identifies a primary metabolic alkalosis, and a PaCO₂ of 47 against a predicted 40 + 0.7 × (34 − 24) = 47 identifies the respiratory compensation as exactly appropriate, so this is a simple rather than a mixed disorder. The anion gap of 7 (7.5 corrected for albumin) excludes a masked anion gap acidosis. The alkalosis is chloride-responsive — chloride 92, generated by contraction and secondary hyperaldosteronism and maintained by chloride and potassium depletion, because pendrin cannot excrete bicarbonate without luminal chloride. Her potassium of 2.9 with a magnesium of 1.4 explains the ECG, and the magnesium explains why the potassium will not correct.
2 · CONNECT. Chain to palpitations: Pharmacology/renal — furosemide inhibits NKCC2, raising distal sodium delivery and flow → renal — collecting duct potassium secretion rises through ROMK and BK → endocrine — volume depletion activates RAAS, and aldosterone amplifies the same secretion → fluid/electrolyte — serum potassium falls to 2.9 → cardiac electrophysiology — reduced extracellular potassium lowers I_K1 conductance (which is proportional to √[K⁺]ₒ), so repolarization slows, the QU interval lengthens, U waves appear, and early afterdepolarizations trigger premature ventricular complexes → cardiovascular — palpitations. Five systems. Chain to weakness: the same first four steps, then neuromuscular — in skeletal muscle, reduced extracellular potassium makes E_K more negative (−107 mV at 2.5 mEq/L), hyperpolarizing the fiber so a larger stimulus is needed to reach threshold → proximal muscle weakness and cramps, compounded by the alkalosis-induced fall in ionized calcium destabilizing sodium channels and by hypomagnesemia. Where they diverge: at the tissue. The same extracellular potassium change makes skeletal muscle less excitable (a gradient effect on E_K) and makes cardiac muscle harder to repolarize (a conductance effect on I_K1). Gradient and conductance are different variables, and that is why one ion produces two opposite-sounding symptoms in the same patient.
3 · PREDICT. Early on, the diuretic dominates and she remains at risk of hypokalemia. As her GFR falls below roughly 30 mL/min/1.73 m², the balance inverts and hyperkalemia becomes the dominant risk, for three converging reasons: fewer functioning nephrons means less total distal sodium delivery and less capacity for potassium secretion; reduced GFR itself limits potassium excretion; and the ACE inhibitor's suppression of aldosterone removes the principal stimulus for collecting duct potassium secretion at exactly the moment that capacity is most needed. Add the type 4 renal tubular acidosis that commonly accompanies diabetic kidney disease, and the same two drugs that produced a potassium of 2.9 today will one day produce a potassium of 6.2. For Chapter 33, a reasonable prediction: that Amara's cardiorenal loop — heart failure reducing renal perfusion, the kidney retaining sodium and water, the retained volume increasing preload on a stiff ventricle — will be run forwards as a single closed model, and that every chapter of this book will be needed to describe one patient.
Review
Level 1 · Recall
27.1 In a 70 kg adult with 42 L of total body water, the plasma volume is approximately:
a) 14 L b) 10.5 L c) 3.5 L d) 28 L
Answer
c — about 3.5 L. Two-thirds of total body water (28 L) is intracellular; one-third (14 L) is extracellular; of the extracellular volume, three-quarters (10.5 L) is interstitial and one-quarter (3.5 L) is plasma. Option (a) is the whole ECF, (b) the interstitium, (d) the ICF. Adding roughly 1.5 L of red cells to the 3.5 L of plasma gives the familiar 5 L blood volume.
27.2 One litre of 5% dextrose in water increases plasma volume by approximately:
a) 1,000 mL b) 250 mL c) 83 mL d) 660 mL
Answer
c — about 83 mL. The dextrose is metabolized within minutes, so what was given is one litre of free water, which distributes across all body water in proportion to compartment size: two thirds (667 mL) intracellular, one third (333 mL) extracellular, and only a quarter of that extracellular share — about 83 mL — remains in plasma. Option (b) is the answer for isotonic saline and (d) for 3% saline. This is why D5W is never used as a volume expander.
27.3 Approximately what percentage of total body potassium is intracellular?
a) 50% b) 75% c) 90% d) 98%
Answer
d — 98%. Of roughly 3,500 mEq of total body potassium, about 3,430 mEq is inside cells and only about 70 mEq is extracellular, of which plasma holds a mere 15–20 mEq. This is why serum potassium responds so dramatically to transcellular shifts and so poorly reflects total body content, and why insulin — which drives potassium into cells — is an effective emergency treatment for hyperkalemia.
27.4 A patient has pH 7.28, PaCO₂ 28 mm Hg, HCO₃⁻ 13 mEq/L. The primary disturbance is:
a) respiratory acidosis b) metabolic acidosis c) respiratory alkalosis d) metabolic alkalosis
Answer
b — metabolic acidosis. The pH is low, so the primary disturbance must be one that lowers pH. The bicarbonate of 13 is low, which lowers pH — that fits. The PaCO₂ of 28 is low, which raises pH, so it cannot be the primary problem and must be compensation. Winter's formula confirms the compensation is appropriate: expected PaCO₂ = 1.5 × 13 + 8 = 27.5 ± 2, and the measured value is 28.
27.5 Winter's formula predicts the expected PaCO₂ in a metabolic acidosis as:
a) 40 + 0.7 × (HCO₃⁻ − 24) b) 1.5 × HCO₃⁻ + 8 ± 2 c) 24 × PaCO₂/HCO₃⁻ d) HCO₃⁻ × 2
Answer
b. Option (a) is the compensation rule for metabolic alkalosis; option (c) is the Henderson equation, used to check a gas for internal consistency rather than to assess compensation. If the measured PaCO₂ is higher than Winter's prediction there is a superimposed respiratory acidosis; if lower, a superimposed respiratory alkalosis.
27.6 In alkalosis, ionized calcium falls because:
a) calcium is excreted more rapidly by the kidney b) H⁺ dissociates from albumin, exposing sites that bind Ca²⁺ c) calcium precipitates with phosphate d) PTH secretion is suppressed
Answer
b. Hydrogen ions and calcium ions compete for the same negatively charged binding sites on albumin. Raising the pH removes hydrogen ions from those sites, exposing more of them, and calcium binds. Total calcium is unchanged — no calcium has left the body — but the ionized, physiologically active fraction falls by roughly 0.05 mmol/L per 0.1 unit rise in pH. This is why acute hyperventilation causes tetany with a completely normal total serum calcium.
27.7 Magnesium depletion causes refractory hypokalemia principally because:
a) magnesium is needed for intestinal potassium absorption b) low intracellular Mg²⁺ removes the block on ROMK, allowing unrestrained renal K⁺ secretion c) magnesium binds potassium in plasma d) hypomagnesemia causes vomiting
Answer
b. Intracellular magnesium normally plugs the ROMK channel pore from the cytoplasmic side, limiting outward (secretory) potassium flux. When intracellular magnesium falls, that block is relieved and the channel conducts potassium freely into the tubular lumen, so administered potassium is secreted straight back out. Two additional contributions: the Na⁺/K⁺ ATPase requires Mg-ATP, so cellular potassium uptake is impaired, and hypomagnesemia mildly stimulates aldosterone.
27.8 The daily fixed (nonvolatile) acid load in a typical adult is approximately:
a) 70 mEq b) 700 mEq c) 15,000 mmol d) 4,300 mEq
Answer
a — about 70 mEq/day (roughly 1 mEq/kg), mostly sulfuric acid from sulfur-containing amino acids and phosphoric acid from phospholipids and phosphoproteins, excreted renally as titratable acid and ammonium. Option (c) is the volatile acid load — the ~15,000 mmol/day of CO₂ exhaled by the lungs, about 200 times larger. Option (d) is the daily filtered bicarbonate load that the kidney must reclaim.
Level 2 · Comprehension
27.9 Explain the statement "sodium determines volume and water determines osmolality," and use it to explain why a patient in decompensated heart failure can be simultaneously sodium-overloaded and hyponatremic.
Model answer
Sodium is the dominant extracellular cation and is actively excluded from cells by the Na⁺/K⁺ ATPase, so sodium and its accompanying anions constitute about 90% of the osmotically active solute holding water in the ECF. Adding sodium to the body therefore expands the ECF; removing it contracts the ECF. Water, by contrast, distributes across all compartments in proportion to their size, so adding water changes the ratio of solute to water — that is, osmolality and hence the serum sodium concentration — while barely changing ECF volume. Two variables, two control systems: RAAS, ANP/BNP, and sympathetic tone regulate sodium (sensing volume and pressure); ADH and thirst regulate water (sensing osmolality).
In decompensated heart failure, the kidney senses a reduced effective arterial blood volume — the heart cannot generate adequate flow despite a full, indeed overfull, circulation. That signal, transmitted through baroreceptors and the juxtaglomerular apparatus, activates RAAS and sympathetic outflow, so the kidney retains sodium avidly and total body sodium climbs; the patient becomes edematous and gains several litres. The same signal also drives non-osmotic ADH release (the baroreceptor override of Figure 31.3), so free water is retained in excess of the sodium. Total body sodium is high; total body water is higher; the ratio falls; the serum sodium is 128. The correct treatment is to remove both — with diuretics and fluid restriction — not to give saline, which would worsen the congestion. The number is a ratio, and you cannot read a numerator from a ratio.
27.10 Explain why the bicarbonate buffer system is effective despite having a pKa of 6.1, which is 1.3 units away from physiological pH.
Model answer
A buffer works best within about one pH unit of its pKa, because that is the range over which both members of the conjugate pair are present in useful quantities. On that criterion, a system with a pKa of 6.1 should be a poor buffer at pH 7.4 — and in a sealed test tube it is.
Two physiological features overturn the chemistry.
It is an open system. The acid member of the pair is carbon dioxide, and the lungs remove it continuously. When acid is added, the reaction H⁺ + HCO₃⁻ → H₂CO₃ → CO₂ + H₂O proceeds to the right and the CO₂ is exhaled rather than accumulating. Because the product is removed, the reaction is never driven back, and the buffer's capacity is effectively unlimited on the acid side. A closed buffer cannot do this: its acid member builds up and its capacity is exhausted.
Both members are independently regulated. The lungs set the CO₂ over minutes; the kidneys set the bicarbonate over hours to days. No other buffer in the body has two dedicated organs adjusting its two components on two timescales. Because pH depends on the ratio, either organ can compensate for a disturbance in the other, and the system can be reset rather than merely resisting change.
The general point is that buffering capacity in a living system is a physiological property, not just a chemical one — and the bicarbonate system is the clearest demonstration in the body.
27.11 A patient has a normal pH of 7.41, a PaCO₂ of 20, and a bicarbonate of 12. Why is "normal pH, no problem" a dangerous conclusion?
Model answer
Because pH reports only the net effect of all processes present, and two opposing severe disorders can cancel exactly.
A PaCO₂ of 20 alone would produce a pH near 7.6; a bicarbonate of 12 alone would produce a pH near 7.2. Both values are far outside their reference ranges and they oppose one another, which means there are two primary disorders, not one disorder with compensation. The compensation rules prove it: if the metabolic acidosis were primary, Winter's formula predicts a PaCO₂ of 1.5 × 12 + 8 = 26 ± 2, and the measured 20 is far below that, indicating a superimposed respiratory alkalosis. If a chronic respiratory alkalosis were primary, the expected bicarbonate would be 24 − 4.5 × 2 = 15, and the measured 12 is below that, indicating a superimposed metabolic acidosis.
The next step is to calculate the anion gap, which will identify the metabolic acidosis as a gap acidosis and point toward the cause. This exact pattern — normal pH, low PaCO₂, low bicarbonate, raised anion gap — is the signature of salicylate toxicity, and it can also appear in sepsis with hepatic dysfunction, or in advanced liver disease with lactic acidosis. The reassurance offered by the pH is precisely inverted: this patient is sicker than one with an obvious acidemia and a single disorder, because two lethal processes are present and neither is visible in the headline number.
27.12 Why does giving insulin to a patient in diabetic ketoacidosis risk a dangerous fall in serum potassium even when the presenting potassium is high?
Model answer
Because the presenting serum potassium reflects a shift, not the total body content, and the total body content is severely depleted.
Three mechanisms have driven potassium out of cells before treatment: insulin deficiency (the Na⁺/K⁺ ATPase is under-stimulated, since insulin normally increases its activity and abundance), hypertonicity from hyperglycemia (water leaves cells, concentrating intracellular potassium and driving it out down its gradient), and to a lesser degree the acidosis. Meanwhile, hours to days of osmotic diuresis have been flushing potassium out of the body: high distal tubular flow and delivery of ketoanions as non-reabsorbable luminal anions both maximize collecting duct potassium secretion, and secondary hyperaldosteronism from volume depletion adds to it. Typical total body deficits are 300–1,000 mEq.
Insulin reverses the shift within minutes, driving potassium back into cells and revealing the true deficit. Serum potassium can fall by 1–2 mEq/L in the first hour, into a range that causes arrhythmia in a patient who is already acidotic, volume-depleted, and hypomagnesemic.
Hence the rule: check the potassium before starting insulin. Below 3.3 mEq/L, give potassium first and delay the insulin. Between 3.3 and 5.3, give potassium with the insulin in the intravenous fluids. Above 5.3, hold potassium and recheck hourly. It is one of the clearest examples in medicine of a laboratory value whose meaning changes entirely once you know which compartment it is reporting on.
Level 3 · Clinical Application
27.13 A 24-year-old marathon runner collapses at mile 24 in warm conditions. She is confused, has vomited twice, and has gained 1.4 kg since the start. Sodium 124 mEq/L. A volunteer starts 1 L of 0.9% saline. Analyze the situation and state what should be done.
Model answer
Diagnosis: symptomatic exercise-associated hyponatremia. The decisive finding is the weight gain — an athlete four hours into a marathon should have lost 1–3 kg, and a gain of 1.4 kg is unambiguous evidence of positive fluid balance. Confusion and vomiting indicate cerebral edema. She is not dehydrated; she is water-overloaded.
Mechanism. She drank hypotonic fluid faster than she could excrete it, and her ability to excrete it was simultaneously impaired by non-osmotic ADH release driven by exercise, pain, nausea, and possibly NSAID use. The osmoreceptor arm of Figure 31.3 was telling her kidney to dump water; the baroreceptor and stress arm overrode it. Sweat sodium losses contributed but are secondary, since sweat is hypotonic and sweating alone raises plasma sodium.
Why the isotonic saline is a mistake. With ADH high, the kidney will retain the free water from the saline while excreting the sodium in a concentrated urine — "desalination" — so a litre of 0.9% saline can lower her serum sodium further and deepen the cerebral edema.
What should be done. Stop the isotonic fluid. Give hypertonic saline: 100 mL of 3% NaCl intravenously over about 10 minutes, repeated up to three times at 10-minute intervals, aiming to raise the serum sodium by 4–5 mEq/L, which is enough to reverse cerebral edema. Restrict oral fluid. Transfer for monitoring, watching for the non-cardiogenic pulmonary edema that accompanies severe cases.
Note the exception to the usual rule. The standard 8 mEq/L per 24 hours limit exists to prevent osmotic demyelination in chronic hyponatremia, where brain cells have extruded organic osmolytes over 24–48 hours. This hyponatremia developed over four hours, so no such adaptation has occurred, there is no demyelination risk, and rapid partial correction is both safe and necessary. Recognizing when a safety rule does not apply requires understanding why it exists.
Prevention: drink to thirst rather than to a schedule.
27.14 A 68-year-old with COPD is admitted with pneumonia. Initial gas on 2 L/min nasal oxygen: pH 7.33, PaCO₂ 62, HCO₃⁻ 32, PaO₂ 58. He is given 15 L/min by non-rebreather mask. One hour later: pH 7.19, PaCO₂ 88, HCO₃⁻ 33, PaO₂ 210, and he is drowsy. Explain both gases and state what went wrong.
Model answer
Gas 1. Acidemia with a raised PaCO₂ → primary respiratory acidosis. Is it acute or chronic? ΔPaCO₂ = 22, i.e. 2.2 "tens." Acute compensation predicts a bicarbonate of 24 + 1 × 2.2 = 26.2 and a pH of 7.40 − 0.08 × 2.2 = 7.22; chronic predicts a bicarbonate of 24 + 3.5 × 2.2 = 31.7 and a pH of 7.40 − 0.03 × 2.2 = 7.33. The measured bicarbonate of 32 and pH of 7.33 match the chronic prediction almost exactly. This is a chronic, well-compensated respiratory acidosis — his baseline — with hypoxemia from the pneumonia. He is not in acute respiratory failure.
Gas 2. PaCO₂ has risen 26 mm Hg in one hour while bicarbonate is essentially unchanged at 33 — exactly as expected, since renal compensation takes days and cannot respond in an hour. The pH has therefore fallen steeply to 7.19, and he is drowsy from CO₂ narcosis. This is an acute respiratory acidosis superimposed on a chronic one, and it was caused by the oxygen.
Why high-flow oxygen raised his PaCO₂, in order of importance: (1) reversal of hypoxic pulmonary vasoconstriction — poorly ventilated lung units, previously vasoconstricted so that perfusion was diverted to better-ventilated regions, now receive blood again, increasing physiological dead space so that the same minute ventilation clears less CO₂; (2) the Haldane effect — oxygenated hemoglobin carries less CO₂ and buffers fewer protons than deoxygenated hemoglobin, so oxygenating the blood displaces CO₂ into the plasma; (3) a modest reduction in hypoxic ventilatory drive, which is real but is the smallest contributor and is often over-stated in teaching.
What went wrong and what to do. The target for a known or suspected CO₂ retainer is an SpO₂ of 88–92%, not maximal oxygenation; a PaO₂ of 210 is far more oxygen than any tissue can use and is doing harm. Titrate the oxygen down to reach that target — but do not abruptly remove it, since rebound hypoxemia can be worse than the starting point. Begin non-invasive ventilation (BiPAP), which directly increases alveolar ventilation and is the definitive treatment for hypercapnic respiratory failure. Treat the pneumonia. Follow the pH, not the PaCO₂ — his baseline PaCO₂ may well be 55–60, and normalizing it would be both unachievable and undesirable.
27.15 A 34-year-old is brought in with tinnitus, nausea, hyperventilation, and confusion. pH 7.38, PaCO₂ 18, HCO₃⁻ 11, Na⁺ 141, Cl⁻ 103, glucose 62, temperature 38.6 °C. Interpret the gas and explain why intubating this patient for airway protection could be fatal.
Model answer
Interpretation. pH 7.38 is technically normal — and misleading. Both the PaCO₂ of 18 and the bicarbonate of 11 are far outside their ranges and oppose one another, so there are two primary disorders. Testing each: if the metabolic acidosis were primary, Winter's formula predicts PaCO₂ = 1.5 × 11 + 8 = 24.5 ± 2, and 18 is well below, indicating a superimposed respiratory alkalosis. If a respiratory alkalosis were primary, the chronic rule predicts a bicarbonate of 24 − 4.5 × 2.2 = 14, and 11 is below that, indicating a superimposed metabolic acidosis. The anion gap is 141 − (103 + 11) = 27, high, confirming a genuine high-anion-gap metabolic acidosis. Delta ratio = (27 − 12)/(24 − 11) = 15/13 = 1.15, a pure gap acidosis.
Diagnosis: salicylate toxicity — mixed primary respiratory alkalosis and primary high-anion-gap metabolic acidosis, with tinnitus, hyperthermia from uncoupled oxidative phosphorylation, and hypoglycemia. Aspirin stimulates the medullary respiratory center directly (the alkalosis) and uncouples oxidative phosphorylation while inhibiting Krebs cycle enzymes, generating lactate and ketones alongside the salicylate anion itself (the acidosis).
Why intubation could kill him. His respiratory alkalosis is not a complication — it is keeping him alive. Salicylic acid has a pKa near 3.0, so at pH 7.4 it is over 99.9% ionized and largely excluded from the central nervous system. As pH falls, the un-ionized fraction rises and salicylate crosses into the brain, where it does its lethal damage. Sedation and paralysis for intubation abolish his compensatory hyperventilation; PaCO₂ rises from 18 toward 40 within minutes; his pH crashes from 7.38 toward 7.0; salicylate floods into the CNS; and patients have arrested at exactly this moment.
What to do instead. Give sodium bicarbonate to raise both serum pH (keeping salicylate ionized and out of the brain) and urine pH above 7.5 (ion-trapping salicylate in the tubule for excretion). Replace potassium, because a hypokalemic kidney will acidify the urine no matter how much bicarbonate is given — the α-intercalated cell H⁺/K⁺-ATPase reclaims potassium at the price of secreting acid. Give glucose regardless of the serum level, since CNS glucose utilization can be impaired at a normal blood concentration and his is already 62. Arrange hemodialysis for his altered mental status. If intubation becomes unavoidable, pre-treat with bicarbonate and set the ventilator to match or exceed his spontaneous minute ventilation through the transition.
Level 4 · Integration and Synthesis
27.16 Construct the complete mechanistic account linking Amara's furosemide prescription to every abnormality on her laboratory sheet — sodium 133, potassium 2.9, chloride 92, bicarbonate 34, BUN 32, creatinine 1.6, magnesium 1.4, pH 7.49, PaCO₂ 47 — and explain why treating her potassium alone would fail.
Model answer
One drug, one transporter, nine abnormalities.
Furosemide inhibits NKCC2 in the thick ascending limb. From that single action:
Sodium 133. Volume depletion triggers the baroreceptor override of osmotic ADH control, so free water is retained in excess of sodium. The kidney is defending volume at the cost of tonicity — a correctly prioritized conflict, not a failure. Reduced diluting capacity from impaired NaCl reabsorption in the diluting segment contributes.
Potassium 2.9. Increased distal sodium delivery (more ENaC flux → more lumen-negative potential), increased tubular flow (steeper chemical gradient plus cilium-mediated BK channel activation), secondary hyperaldosteronism from volume depletion, alkalosis-driven intracellular shift, and increased distal bicarbonate delivery acting as a non-reabsorbable anion. Five mechanisms, four of them downstream of the drug's site of action.
Chloride 92. The loop of Henle is the principal site of chloride reabsorption. Blocking NKCC2 wastes chloride obligatorily.
Bicarbonate 34 and pH 7.49. Metabolic alkalosis, generated by contraction (chloride-rich, bicarbonate-poor fluid lost), by aldosterone-driven distal H⁺ secretion, and by hypokalemia-driven intracellular H⁺ shift with upregulated H⁺/K⁺-ATPase and ammoniagenesis; maintained by chloride and potassium depletion, since pendrin in type B intercalated cells cannot excrete bicarbonate without luminal chloride to exchange.
PaCO₂ 47. Appropriate respiratory compensation: expected 40 + 0.7 × 10 = 47. Not a disorder.
BUN 32 and creatinine 1.6, ratio 20. Prerenal azotemia. Volume depletion reduces renal perfusion; the proximal tubule reabsorbs more sodium and water and, with it, more urea, raising the BUN disproportionately. Her lisinopril adds a hemodynamic component by dilating the efferent arteriole and lowering filtration pressure, and her pre-existing chronic kidney disease reduces the reserve to absorb any of it.
Magnesium 1.4. The lumen-positive potential of the thick ascending limb drives paracellular magnesium reabsorption through claudin-16/19. Abolishing that potential wastes magnesium.
Why treating the potassium alone fails. Three reasons, and they are independent. First, the magnesium: low intracellular Mg²⁺ removes the physiological block on ROMK, so administered potassium is secreted straight back into the urine, and Mg-ATP deficiency impairs cellular uptake. Second, the chloride: without chloride, pendrin cannot excrete the excess bicarbonate, so the alkalosis persists — and the alkalosis keeps shifting potassium into cells and keeps delivering bicarbonate distally as a non-reabsorbable anion, both of which perpetuate the hypokalemia. Third, the volume depletion: while RAAS remains activated, aldosterone continues to drive potassium secretion regardless of how much you replace.
Therefore the correct treatment attacks all three simultaneously: magnesium first or concurrently, potassium given as chloride (repairing potassium and chloride together), cautious isotonic volume repletion, and reduction of the furosemide dose — with consideration of spironolactone, which would block the aldosterone arm at its source and has independent mortality benefit in heart failure.
27.17 Compare the acid-base and electrolyte consequences of severe vomiting with those of severe diarrhea. Both lose gastrointestinal fluid and both cause hypokalemia — yet one causes alkalosis and the other acidosis. Explain fully, and explain paradoxical aciduria.
Model answer
The determining variable is which secretion is lost, not how much.
Vomiting removes gastric fluid: H⁺ 60–100 mEq/L, Cl⁻ 100–140 mEq/L, K⁺ only 5–10 mEq/L. Every mEq of HCl secreted by parietal cells leaves an mEq of HCO₃⁻ in the blood — the alkaline tide, normally cancelled when duodenal contents are neutralized by pancreatic bicarbonate. Vomiting removes the acid before that cancellation, so the bicarbonate stands unopposed. Result: metabolic alkalosis with hypochloremia.
Diarrhea removes intestinal and pancreatic fluid: HCO₃⁻ 30–70 mEq/L, and colonic fluid potassium up to 30–90 mEq/L. Losing bicarbonate is chemically equivalent to gaining acid, and chloride is retained to preserve electroneutrality. Result: normal-anion-gap hyperchloremic metabolic acidosis.
Both cause hypokalemia, by different routes. In diarrhea it is a direct loss — the fluid itself is potassium-rich. In vomiting it is almost entirely renal: gastric fluid contains little potassium, but the alkalosis delivers bicarbonate distally as a non-reabsorbable anion (deepening lumen negativity) and the volume depletion drives aldosterone. The stomach starts it; the kidney does the losing.
Paradoxical aciduria. A patient with an arterial pH of 7.55 from prolonged vomiting produces acid urine, which appears to be the opposite of what regulation should do. It is not a failure. The kidney has been asked to defend three variables at once and they conflict:
- Volume is low → RAAS is activated → aldosterone drives distal H⁺ and K⁺ secretion, and the proximal tubule avidly reabsorbs sodium with bicarbonate.
- Chloride is low → pendrin cannot secrete bicarbonate.
- Potassium is low → H⁺ shifts into cells, the H⁺/K⁺-ATPase reclaims potassium at the price of secreting acid, and ammoniagenesis rises.
- pH is high → regulation would call for bicarbonate excretion.
Volume wins, because volume failure kills in minutes while alkalemia takes much longer. The kidney therefore secretes acid into the urine while the patient is alkalemic — precisely the prioritized-conflict pattern that also explains hypovolemic hyponatremia. It is the same principle as Chapter 1's observation that Amara's compensations were making her heart worse: homeostatic loops are not context-aware, and when two of them conflict, the more urgent variable is defended at the other's expense.
Treatment follows directly. Vomiting: sodium chloride and potassium chloride — restore volume and supply the chloride that permits bicarbonate excretion. Diarrhea: volume and bicarbonate (or lactate/acetate) replacement plus potassium — which is exactly the composition of oral rehydration solution, whose glucose additionally drives sodium absorption through SGLT1 in a gut whose other transport is impaired.
27.18 A patient arrives with pH 7.09, PaCO₂ 26, HCO₃⁻ 8, Na⁺ 136, Cl⁻ 100, K⁺ 5.9, glucose 88, lactate 2.1, albumin 2.0 g/dL, and an osmolar gap of 24 mOsm/kg. Work the full five steps, identify what is present, and explain what the albumin and the osmolar gap add.
Model answer
Step 0 · Consistency. [H⁺] = 24 × 26/8 = 78 nmol/L → pH ≈ 7.11, consistent with 7.09.
Step 1 · pH 7.09 → severe ACIDEMIA.
Step 2 · Primary disturbance. HCO₃⁻ 8 is low → metabolic acidosis. PaCO₂ 26 is low, which raises pH, so it is compensation.
Step 3 · Compensation. Winter's: expected PaCO₂ = 1.5 × 8 + 8 = 20 (± 2, so 18–22). Measured 26 — higher than expected, so respiratory compensation is inadequate: there is a superimposed respiratory acidosis. Clinically this means the patient is tiring, has been sedated, or has a coexisting pulmonary problem, and it is an ominous finding that argues for prompt airway support.
Step 4 · Anion gap = 136 − (100 + 8) = 28, high. But the albumin is 2.0, and albumin is the principal unmeasured anion contributing to the normal gap. Correction: add 2.5 for every 1 g/dL below 4.0, so corrected AG = 28 + 2.5 × 2.0 = 33 — even higher than it first appears. The albumin correction matters in the opposite direction too: a hypoalbuminemic patient with a "normal" gap of 7 may actually have a significant gap acidosis being concealed.
Step 5 · Delta ratio = (28 − 12)/(24 − 8) = 16/16 = 1.0 using the uncorrected gap, or (33 − 12)/16 = 1.3 corrected. Both are in the 1.0–2.0 range → a pure anion gap acidosis, with no additional non-gap acidosis and no coexisting metabolic alkalosis.
What the osmolar gap adds — and it is the key to the case. An osmolar gap of 24 mOsm/kg (normal < 10) means there are unmeasured osmotically active molecules in the plasma. Combined with a large anion gap that is not explained by the usual suspects — glucose is 88, so no ketoacidosis; lactate is 2.1, essentially normal; no history of renal failure is implied — this combination points strongly to a toxic alcohol ingestion: methanol or ethylene glycol.
The mechanism explains both gaps at once. The parent alcohol is a small, uncharged molecule that raises measured osmolality without contributing to the anion gap. Alcohol dehydrogenase then metabolizes it to a charged organic acid — formic acid from methanol, glycolic and oxalic acid from ethylene glycol — which raises the anion gap while lowering the osmolar gap as the parent compound is consumed. So early presentation shows a large osmolar gap and a small anion gap; late presentation shows the reverse; and this patient, with both raised, is in between. Ethanol also raises the osmolar gap and should be excluded.
Complete diagnosis: severe high-anion-gap metabolic acidosis (corrected gap 33) with inadequate respiratory compensation — that is, a superimposed respiratory acidosis — an osmolar gap of 24, and hyperkalemia from the acidosis-driven transcellular shift; the pattern of a toxic alcohol ingestion.
Treatment logic: block alcohol dehydrogenase with fomepizole (or ethanol) to prevent further conversion of the parent alcohol to its toxic acid metabolite; give bicarbonate for the severe acidemia; arrange hemodialysis, which removes both the parent alcohol and its metabolites; give the specific cofactors — folate for methanol, thiamine and pyridoxine for ethylene glycol — that shunt the metabolites toward harmless products; and secure the airway, since the inadequate respiratory compensation predicts imminent decompensation. Note that treating the potassium of 5.9 as a primary problem would be an error: it is a shift, and it will fall as the acidosis is corrected — indeed it may fall too far, since total body potassium in a patient with an acute organic acidosis is usually normal or low.
Concept Map to Complete
Copy this onto blank paper and complete every bracket from memory before checking the chapter. Then, in a second color, add what you missed.
TOTAL BODY WATER = [ __ ]% of body mass
│ (adult male)
┌────────────────┴─────────────────┐
ICF = [ __ ] ECF = [ __ ]
= [ __ ] L in a 70 kg adult = [ __ ] L
dominant cation [ __ ] dominant cation [ __ ]
│ ┌─────────┴─────────┐
│ INTERSTITIAL PLASMA
│ [ ___ ] L [ ___ ] L
│ └── separated by [ _______ ]
│ pressure ≈ [ __ ] mm Hg
└───────── maintained by [ ______________ ] ─────────┘
(3 __ out / 2 __ in per ATP)
═══ TWO SEPARATE VARIABLES, TWO SEPARATE CONTROL SYSTEMS ═══
[ ______ ] determines ECF VOLUME [ _____ ] determines OSMOLALITY
controlled by RAAS, [ ___ ]/BNP controlled by [ ___ ] and thirst
sensed by [ ______ ] receptors sensed by [ ____________ ]
│ │
└──────────────┬───────────────────┘
▼
SERUM Na+ CONCENTRATION tells you
[ nothing / everything ] about total body Na+
═══════════════ ACID-BASE ═══════════════
volatile acid = [ ____ ] ≈ [ ______ ] mmol/day → removed by [ _____ ]
fixed acid ≈ [ __ ] mEq/day → removed by [ _______ ]
pH depends on the [ _____ ] of HCO3- to CO2 = [ __ ] : 1
│
┌──────────────────────────┼──────────────────────────┐
BUFFERS RESPIRATORY RENAL
[ ______ ] [ ______ ] [ ______ ]
(time) (time) (time)
│
FIVE STEPS TO READ ANY BLOOD GAS
1 [ ____ ] 2 [ ______________ ] 3 [ ____________ ]
4 [ __________ ] 5 [ ___________ ]
Winter's formula: expected PaCO2 = [ ___ ] × HCO3- + [ _ ] ± 2
Anion gap = [ ___ ] − ( [ ___ ] + [ _____ ] ) normal [ _ ]–[ __ ]
Lab / Self-Exploration
- Calculate your own compartments. Take your body mass in kilograms. Estimate total body water as 60% (male) or 52% (female); then ICF as two-thirds and ECF as one-third of that, and plasma as one-quarter of the ECF. Compare your calculated plasma volume with the 450–500 mL taken in a blood donation — what fraction of your plasma volume is that, and why do you feel it?
- Read a bag. Find the label of any intravenous fluid bag (a photograph online will do) and list its sodium, chloride, potassium, buffer, and osmolarity. Then predict, using Figure 31.2, where one litre of it would go in a 70 kg adult. Do this for normal saline, lactated Ringer's, and D5W and compare.
- Work a gas from memory. Cover the answers to the seven worked examples in §31.8 and redo each one using only the five steps. Time yourself; a fluent clinician does this in under ninety seconds. Then invent three gases of your own and check them with the Henderson equation — you will find it is surprisingly hard to invent an internally consistent one, which is exactly why the check works.
- Watch respiratory compensation happen. Sit quietly for two minutes, then breathe deeply and rapidly for no more than 30 seconds (stop immediately if you feel light-headed, and do this seated, never standing or in water). Note any tingling around the mouth or in the fingers. You have just lowered your PaCO₂, raised your pH, and reduced your ionized calcium by the albumin-binding mechanism of §31.5 — the fastest homeostatic demonstration in this book. Note also how quickly the sensation resolves once you stop.
- Estimate your sweat sodium loss. Weigh yourself, unclothed and dry, before and after an hour of exercise, recording anything you drink. Sweat volume ≈ (weight lost in kg) + (fluid drunk in kg). Multiply by 40 mEq/L to estimate sodium lost, and convert to grams of salt (1 mEq Na⁺ ≈ 58.5 mg of NaCl). Compare with the sodium content of whatever you would normally drink afterwards.
- Audit a food label for sodium. Convert the milligrams of sodium on any packaged food to milliequivalents (divide by 23). Then work out how many litres of extracellular fluid that sodium would hold if fully retained, given an ECF sodium concentration of 140 mEq/L. The number is smaller than people expect for one meal and larger than they expect for a day.
Key Terms
aldosterone · Adrenal mineralocorticoid that increases ENaC, Na⁺/K⁺ ATPase, and ROMK in collecting duct principal cells, raising sodium reabsorption and potassium secretion.
anion gap · Na⁺ − (Cl⁻ + HCO₃⁻), normally 8–12 mEq/L; raised when an unmeasured anion accompanies a metabolic acidosis. Must be corrected for albumin.
antidiuretic hormone (ADH, vasopressin) · Posterior pituitary hormone that inserts aquaporin-2 into collecting duct principal cells, concentrating urine; released by rising osmolality and, more powerfully, by falling volume.
aquaporin-2 · The ADH-regulated water channel of the collecting duct apical membrane.
chloride-responsive alkalosis · Metabolic alkalosis with a urine chloride below 20 mEq/L, maintained by chloride depletion and corrected by giving sodium and potassium chloride.
contraction alkalosis · Metabolic alkalosis produced when chloride-rich, bicarbonate-poor fluid is lost, leaving the same bicarbonate in a smaller extracellular volume.
delta-delta (delta ratio) · (AG − 12)/(24 − HCO₃⁻); identifies a mixed metabolic disorder hiding behind an anion gap acidosis.
diffusion trapping · The mechanism by which lipid-soluble NH₃ enters the collecting duct lumen, combines with secreted H⁺, and becomes charged NH₄⁺ that cannot diffuse back out.
effective arterial blood volume · The portion of arterial volume actually perfusing tissue and sensed by baroreceptors; low in heart failure and cirrhosis despite total volume overload.
ENaC · The amiloride-sensitive epithelial sodium channel of the collecting duct principal cell apical membrane; the target of aldosterone.
extracellular fluid (ECF) · One-third of total body water (~14 L in a 70 kg adult), divided into interstitial fluid (~10.5 L) and plasma (~3.5 L).
fixed (nonvolatile) acid · Sulfuric, phosphoric, and organic acids, ~70 mEq/day, removable only by the kidney.
Henderson equation · [H⁺] (nmol/L) = 24 × PaCO₂/[HCO₃⁻]; used to check whether a blood gas is internally consistent.
Henderson-Hasselbalch equation · pH = 6.1 + log([HCO₃⁻]/(0.03 × PaCO₂)); formally states that pH depends on the ratio of bicarbonate to dissolved carbon dioxide.
hypernatremia · Serum Na⁺ > 145 mEq/L; always a water deficit relative to sodium, and almost always requiring impaired thirst or impaired access to water.
hyponatremia · Serum Na⁺ < 135 mEq/L; classified first by tonicity, then by volume status.
intercalated cells (type A and type B) · Dark, mitochondria-rich collecting duct cells; type A secretes H⁺ and returns HCO₃⁻ to the blood, type B is its mirror image and secretes HCO₃⁻ through pendrin.
intracellular fluid (ICF) · Two-thirds of total body water (~28 L in a 70 kg adult), potassium- and phosphate-dominated.
isohydric shift · The coupling by which deoxyhemoglobin, a better proton acceptor than oxyhemoglobin, buffers the H⁺ generated by CO₂ hydration in the tissues.
metabolic water · Water produced by oxidative metabolism, ~250–350 mL/day.
obligatory water loss · The ~1,400 mL/day that cannot be reduced: insensible losses, fecal water, and the minimum urine volume set by solute load divided by maximum urine concentration.
osmolality · Osmoles per kilogram of solvent; normal plasma 275–295 mOsm/kg.
osmolar gap · Measured minus calculated osmolality; above 10 mOsm/kg indicates unmeasured osmoles such as ethanol, methanol, ethylene glycol, or mannitol.
osmotic demyelination syndrome · Delayed demyelinating injury, classically pontine, caused by correcting chronic hyponatremia faster than about 8 mEq/L per 24 hours.
pendrin · The apical Cl⁻/HCO₃⁻ exchanger of type B intercalated cells; requires luminal chloride, which is why alkalosis is chloride-responsive.
physiological reserve · See Chapter 30; in this chapter it is why elderly patients develop every disorder described here more readily than young ones.
principal cell · Pale collecting duct cell carrying ENaC, ROMK, BK, and aquaporin-2; the target of aldosterone and ADH.
refeeding syndrome · The fall in serum phosphate, potassium, and magnesium that follows reintroduction of nutrition after starvation, driven by an insulin-mediated shift into cells.
ROMK · The renal outer medullary potassium channel of the principal cell apical membrane; blocked from the cytoplasmic side by intracellular magnesium.
titratable acid · Acid excreted bound to urinary buffers, principally phosphate; 10–40 mEq/day, and the route by which the kidney generates some new bicarbonate.
tonicity (effective osmolality) · The part of osmolality contributed by solutes that cannot freely cross cell membranes, and therefore the only part that moves water; 2 × Na⁺ + glucose/18.
transcellular fluid · The 1–2 L of cerebrospinal, synovial, pleural, pericardial, peritoneal, aqueous, and gastrointestinal fluid.
volatile acid · Carbon dioxide, ~15,000 mmol/day, removable by the lungs.
Winter's formula · Expected PaCO₂ = 1.5 × [HCO₃⁻] + 8 ± 2; the respiratory compensation rule for metabolic acidosis.
Next: Chapter 32 · Integrated Exercise Physiology — where every balance struck in this chapter is struck again under maximal load, and where Amara's second cardiopulmonary exercise test measures how much of her physiology eighteen months bought back.