Acid-base: buffers & compensation

Why this matters to the anaesthetist

The arterial blood gas is the most frequently ordered perioperative test, and the disorders it reveals are the everyday currency of anaesthetic and intensive-care practice. Under anaesthesia the patient apnoeas, is ventilated, becomes hypothermic, is given fluids and opioids, and is subjected to shock and reperfusion, and each of these deranges acid-base balance. A respiratory acidosis from permissive hypercapnia, a metabolic acidosis from shock and lactate, and a metabolic alkalosis from gastric drainage are read off the same three numbers, pH, PaCO2 and bicarbonate. The skill is to convert those three numbers into a named disorder, to check whether the expected compensation is present, and to act on the cause rather than on the number [1][2][4].

pH and the hydrogen ion — the defended variable

Arterial pH is normally 7.35 to 7.45, a vanishingly narrow range that corresponds to a hydrogen-ion concentration of only 35 to 45 nmol per litre (mean 40 nmol per litre at pH 7.40). The body defends this range because protein structure and enzyme function, potassium distribution, and the binding of catecholamines and drugs to receptors are all exquisitely sensitive to it. To put the scale in perspective, the daily acid load the body handles is enormous, roughly 15000 mmol of volatile acid (carbon dioxide) excreted by the lungs and about 70 mmol of fixed (non-volatile) acid excreted by the kidneys, against a free extracellular hydrogen-ion pool of well under one millimole. The whole system exists to keep that tiny pool constant [4][12].

pH is the negative logarithm of the hydrogen-ion concentration, and the logarithm has two consequences the candidate must carry. First, each 0.3 unit fall in pH roughly doubles the hydrogen-ion concentration and each 0.3 unit rise halves it: pH 7.40 is 40 nmol per litre, 7.10 is about 80, 7.00 is about 100, and 7.70 is about 20. Second, the relationship is curved, not linear, so judging severity from the pH alone understates the acid load at low pH. The clinically useful linear form, which avoids the logarithm, is the Henderson equation: hydrogen ion (in nmol per litre) equals 24 times PaCO2 (in mmHg) divided by bicarbonate (in mmol per litre). At PaCO2 40 and bicarbonate 24 this gives 40 nmol per litre, the value at pH 7.40 [1].

What a buffer is — the weak acid and its conjugate base

A buffer is a mixture of a weak acid and its conjugate base (or a weak base and its conjugate acid) that resists a change in pH when acid or base is added. The weak acid is denoted HA and dissociates into its conjugate base A minus and a hydrogen ion. When a strong acid is added, the surplus hydrogen ions are taken up by A minus to form more HA, so the pH moves only a little; when a strong base is added, it is neutralised by HA, which releases its hydrogen ion and is converted to A minus. A buffer therefore buffers against added acid through its conjugate-base component and against added base through its weak-acid component [5].

Three properties define a buffer and are worth stating precisely. The pKa is the pH at which the weak acid is half dissociated, so that HA equals A minus and the ratio is 1. A buffer is most effective within about one pH unit either side of its pKa, because that is where both components are present in appreciable amounts. The buffering capacity (or buffer value, beta) is the amount of acid or base, in mmol per litre, needed to shift the pH by one unit; it is greatest at the pKa and falls away symmetrically. The concentration of the buffer matters as much as its pKa, because buffering capacity is proportional to the total amount of buffer present. A useful mnemonic is that a buffer does most of its work near its pKa, in proportion to its concentration, and only over roughly two pH units [5].

The Henderson-Hasselbalch equation

The Henderson-Hasselbalch equation is the algebraic rearrangement of the dissociation expression and is the single equation on which all of clinical acid-base rests [1][12]:

• pH equals pKa plus the log of the ratio of the conjugate base to the weak acid. [1]

For the bicarbonate buffer the weak acid is carbonic acid and the conjugate base is bicarbonate, but because carbonic acid is in rapid equilibrium with dissolved CO2, the practical form is written as: [1]

• pH equals 6.1 plus the log of (bicarbonate divided by (0.03 times PaCO2)). [1]

Two deductions follow immediately and are the heart of the topic. First, pH is set by the ratio, not by either component alone. A low bicarbonate with a proportionately low PaCO2 gives a normal pH; a normal bicarbonate with a high PaCO2 gives an acidosis. Second, because PaCO2 is exhaled by the lung and bicarbonate is regenerated by the kidney, either organ can move the ratio and defend the pH. This is the physiological basis of compensation: a metabolic disorder (a bicarbonate problem) is compensated by changing PaCO2, and a respiratory disorder (a PaCO2 problem) is compensated by changing bicarbonate [1][7].

The bicarbonate buffer — why a non-ideal pKa dominates

The bicarbonate buffer has a pKa of 6.1, which is a long way below the physiological pH of 7.40. By the rule that a buffer works best within one pH unit of its pKa, bicarbonate should be a poor extracellular buffer, and a closed bicarbonate buffer at pH 7.40 would indeed be near-useless. The reason it dominates is that it is open: the acid component (CO2) is excreted by the lung and the base component (bicarbonate) is regenerated by the kidney, so neither side of the ratio is fixed. The system is therefore not a sealed flask being titrated but a throughput, with the two excretory organs holding the ratio steady against a colossal daily acid load [1][7].

Three consequences of the open design deserve emphasis. First, the buffering capacity is amplified without limit in principle, because the lung can exhale any amount of CO2 the tissues generate. Second, the pKa of 6.1 makes the system a near-perfect sensor of ventilation: because the ratio of bicarbonate to dissolved CO2 at pH 7.40 is about 20 to 1 (24 over 1.2, since 0.03 times 40 is 1.2), a small change in PaCO2 produces a readily measurable change in pH, which is exactly what a ventilatory controller needs. Third, the lung is the fast responder and the kidney the slow one: CO2 is exhaled in seconds to minutes, whereas renal bicarbonate regeneration takes hours to days, and this asymmetry in speed is the single most important fact about compensation [1][7].

The 20 to 1 ratio at pH 7.40 is worth memorising because it explains both the sensitivity of pH to PaCO2 and the geometry of the Davenport diagram (below). Putting the numbers into the equation: 6.1 plus the log of 20 equals 6.1 plus 1.3, which is 7.40. Log 2 is 0.3, so a doubling of either component shifts the pH by 0.3 in the opposite direction [1].

The physiological buffers compared

No single buffer handles the whole acid load; the body runs several in parallel, and because they all equilibrate with the same hydrogen-ion concentration they behave as a single shared system (the isohydric principle). The five systems the anaesthetist should know are compared below [5][6].

[1]

The isohydric principle and the buffer base

Because every buffer pair in a solution shares the same hydrogen-ion concentration, they all equilibrate together: this is the isohydric principle, and it means that a change in the ratio of any one buffer is reflected in a corresponding change in all the others. The bicarbonate ratio and the haemoglobin ratio move together. The practical consequence is that the non-bicarbonate buffers (haemoglobin, plasma proteins, phosphate) collectively constitute a fixed quantity called the buffer base, and the total concentration of buffer in whole blood is roughly 45 to 50 mmol per litre. When an acid is added, the bicarbonate fraction falls and the non-bicarbonate buffers take up some of the load; the proportions shift but the total buffer base is unchanged until the kidney regenerates bicarbonate. This is the conceptual basis of base excess (below), which quantifies how much acid or base must be added to return the blood to a standard state [3][5].

The Haldane effect — why deoxyhaemoglobin is the better buffer

The Haldane effect is the most examinable detail of the haemoglobin buffer and is inseparable from carbon dioxide transport. Haemoglobin carries about 20 per cent of the venous CO2 as carbamino compounds (bound to terminal amino groups), and it buffers the hydrogen ions released when CO2 is hydrated to carbonic acid, chiefly through its imidazole side chains on histidine. Deoxygenated haemoglobin is a better buffer than oxygenated haemoglobin because removal of oxygen shifts the pKa of these groups upward, increasing their ability to take up a hydrogen ion [6].

The physiological payoff is elegant. In the tissues, where haemoglobin gives up oxygen and becomes deoxygenated, its buffering capacity rises just as CO2 is being generated, so CO2 is loaded and buffered with ease. In the lungs, oxygenation converts haemoglobin back to the more acidic form, releasing hydrogen ions that combine with bicarbonate to reform CO2 for excretion. The Haldane effect is therefore the molecular reason venous blood can carry more CO2 than arterial blood at the same PCO2, and it accounts for roughly a third of CO2 transport. The anaesthetic relevance is direct: anything that reduces haemoglobin (anaemia, haemodilution) or that changes its oxygenation state shifts the buffering capacity of whole blood [6].

The four primary disorders

There are four primary acid-base disorders, defined by which variable is primarily abnormal. A respiratory disorder is a PaCO2 problem: PaCO2 rises in respiratory acidosis (hypoventilation) and falls in respiratory alkalosis (hyperventilation). A metabolic disorder is a bicarbonate problem: bicarbonate falls in metabolic acidosis (acid gain or bicarbonate loss) and rises in metabolic alkalosis (acid loss or bicarbonate gain). Each primary disorder evokes a compensatory response from the opposite system, always in the direction that returns the pH towards (but never beyond) normal [1][2].

A respiratory disorder is, mechanically, a change in alveolar ventilation relative to CO2 production; a metabolic disorder is a change in the extracellular bicarbonate concentration from addition or loss of acid or base. The two classes are kept separate because their compensations are governed by different rules and operate on different timescales, and because mixing them is the rule rather than the exception in sick patients [1][2].

Compensation — respiratory is fast, renal is slow

The body compensates for a primary disorder by moving the opposite variable, and the cardinal fact is the asymmetry of speed. Respiratory compensation is fast: alveolar ventilation can change within seconds and is essentially complete within minutes to a few hours, because the medullary chemoreceptors respond briskly to the hydrogen-ion concentration of the CSF (which tracks PaCO2). Renal (metabolic) compensation is slow: the kidney must reabsorb or excrete bicarbonate, generate new bicarbonate by ammonium and titratable-acid excretion, and these processes take two to five days to reach steady state [7][8].

[1]

The speed asymmetry generates three clinically load-bearing rules. An acute respiratory disorder shows little compensation, because the kidney has not yet had time to act; a chronic respiratory disorder is well compensated, because it has. A metabolic disorder develops its respiratory compensation within hours, so even a short-standing metabolic acidosis usually has a compensatory fall in PaCO2. Full renal compensation of a respiratory disorder takes days, which is why the same PaCO2 produces a much lower pH in acute respiratory acidosis than in chronic [7][8].

Respiratory compensation for metabolic disorders

For a metabolic disorder, the lung adjusts PaCO2. In metabolic acidosis, the fall in pH stimulates the medullary chemoreceptors and produces the deep, sighing Kussmaul ventilation of, classically, diabetic ketoacidosis; PaCO2 is driven down until the Henderson ratio is partly restored. The expected PaCO2 is given by Winter's formula: expected PaCO2 equals 1.5 times the bicarbonate plus 8, plus or minus 2 mmHg [9]. A measured PaCO2 within this band means the respiratory compensation is appropriate; a PaCO2 higher than expected means a coexisting respiratory acidosis (the patient is also hypoventilating), and a PaCO2 lower than expected means a coexisting respiratory alkalosis (additional hyperventilation from pain, hypoxia or sepsis). Failure to apply Winter's formula is the commonest single error in acid-base interpretation [1][9].

In metabolic alkalosis, the raised pH should depress ventilation and raise PaCO2, and this does occur, but the compensation is limited by hypoxia: as PaCO2 rises, alveolar oxygen falls, and hypoxic drive eventually constrains further hypoventilation. The PaCO2 therefore rarely rises above about 55 mmHg, even in severe alkalosis. A rough expectation is that PaCO2 rises about 0.7 mmHg for each 1 mmol per litre rise in bicarbonate, but the more robust rule is the ceiling: a PaCO2 above 55 to 60 mmHg in a metabolic alkalosis signals coexisting respiratory disease rather than pure compensation [2].

Renal compensation for respiratory disorders

For a respiratory disorder, the kidney adjusts bicarbonate. The renal response has two components and two timescales, and the magnitude of the chronic response is what defines the compensation rules [7][8].

[1]

The mechanism of renal compensation runs along the nephron and is itself examinable. The proximal tubule reabsorbs about 85 per cent of the filtered bicarbonate through the NHE3 sodium-hydrogen antiporter (sodium in, hydrogen out) and luminal carbonic anhydrase: secreted hydrogen combines with filtered bicarbonate to form carbonic acid, which is dehydrated to CO2 and water by carbonic anhydrase IV on the luminal membrane, the CO2 diffusing back into the cell where carbonic anhydrase II rehydrates it; the hydrogen is re-secreted and the bicarbonate exits basolaterally on the Na-bicarbonate cotransporter. In respiratory acidosis the kidney upregulates NHE3 and the Na-bicarbonate cotransporter to retain more bicarbonate; in respiratory alkalosis it downregulates them to excrete it [7][8].

The distal nephron, especially the type A intercalated cell of the collecting duct, secretes hydrogen via the H-ATPase (and the H-K-ATPase), and can drive the urine down to a pH of about 4.5, the limiting minimal urine pH. Beyond simply reabsorbing filtered bicarbonate, the kidney generates new bicarbonate by two routes. The first is ammonium (NH4+) excretion: proximal glutamine metabolism yields ammonium and alpha-ketoglutarate, the latter being metabolised to bicarbonate that is returned to the blood; ammonium is secreted into the lumen (in place of hydrogen on the NHE3 transporter, or by NH3 diffusion and trapping) and excreted. This is the adaptive response to chronic acidosis, and it can rise from about 40 to over 250 mmol per day, making it the dominant route for excreting a fixed acid load [8]. The second is titratable acid, chiefly phosphate (HPO4 2 minus to H2PO4 minus), contributing about 30 mmol per day. The anaesthetic point is that the kidney's ability to compensate respiratory disorders is powerful but slow, and that the chronic respiratory rules are a direct read-out of ammonium adaptation [7][8].

The compensation rules in one place

The four compensation rules should be committed to memory as a single block, because they are the arithmetic backbone of every acid-base interpretation [1][2][8].

[1]

Mixed disorders

A single acid-base disturbance with appropriate compensation is the teaching case; the reality in sick patients is that two or more primary disorders coexist, and the compensation rules are precisely the tool that unmasks them. The principle is to compute the expected compensation and then compare it with the measured value: a measured value that falls outside the predicted band means a second primary disorder is present [1][2].

Several mixed patterns recur and should be recognised. A metabolic acidosis with respiratory acidosis (the winter COPD patient with a pneumonia, or the overdose with both hypoventilation and a lactic acidosis) gives a PaCO2 higher than Winter predicts and a pH that is very low. A metabolic acidosis with respiratory alkalosis (sepsis, salicylate toxicity, hepatorenal failure) gives a PaCO2 lower than Winter predicts. A metabolic alkalosis with respiratory acidosis (the COPD patient on diuretics, or chronic vomiting with hypoventilation) is the classic near-normalising mixture, where the pH may even be normal because the two disorders pull in opposite directions. Crucially, compensation never overcorrects the pH: a normal pH with an abnormal PaCO2 and bicarbonate always means a mixed disorder, never a fully-compensated single one [1][2].

For metabolic acidosis, the delta-delta (delta ratio) refines the analysis. The delta anion gap (the observed gap minus 12) is compared with the delta bicarbonate (24 minus the observed bicarbonate). A ratio of 1 to 2 is a pure high-anion-gap acidosis; a ratio below 1 means a coexisting normal-anion-gap (hyperchloraemic) acidosis, because bicarbonate has fallen more than the anion gap has risen; a ratio above 2 means a coexisting metabolic alkalosis, because bicarbonate has fallen less than expected. Salicylate toxicity classically gives a high ratio with a mixed high-gap acidosis and respiratory alkalosis [1][10].

Base excess and standard bicarbonate

Base excess (or base deficit when negative) is the amount of acid or base, in mmol per litre, that must be added to fully oxygenated blood to return it to a pH of 7.40 at a PaCO2 of 40 mmHg and 37 degrees Celsius. It is, by construction, the metabolic component of an acid-base disorder measured at a standard PaCO2, and so it is independent of the current respiratory status: a patient with a respiratory acidosis and an appropriately raised bicarbonate has a near-normal base excess, whereas a patient with a primary metabolic acidosis has a negative base excess regardless of PaCO2 [3][11].

The normal base excess is plus or minus 2 mmol per litre. A base excess below minus 2 is a metabolic acidosis (a base deficit); a base excess above plus 2 is a metabolic alkalosis. Two methodological points matter. First, the standard bicarbonate (the bicarbonate concentration that blood would have at PaCO2 40) carries the same information in a different form, and both are reported by every modern blood-gas analyser; the standard bicarbonate is normal 24 mmol per litre. Second, base excess can be calculated for whole blood or for the extracellular fluid (the in-vivo or standard base excess), and the two differ because haemoglobin distributes differently between red cells and the interstitial space; the extracellular-fluid value is the more clinically useful, and Lang and Zander showed that the calculated value is accurate to within experimental error when the haemoglobin is known [3][11].

The anion gap

The anion gap is the difference between the routinely measured cations and anions, conventionally sodium minus (chloride plus bicarbonate), and it is normally 8 to 12 mmol per litre (12 to 16 by some laboratories). It exists because plasma contains unmeasured anions (proteins, phosphate, sulphate, organic anions) that balance the unmeasured cations (potassium, calcium, magnesium) with a small surplus on the anion side. Its usefulness is to separate the metabolic acidoses into two mechanistic families [1][6].

A high anion-gap metabolic acidosis means an unmeasured acid has been added: the acid consumes bicarbonate and is replaced by its anion, which widens the gap. The causes are remembered as MUDPILES (methanol, uraemia, diabetic ketoacidosis, propylene glycol, iron or isoniazid, lactic acidosis, ethylene glycol, salicylates) or GOLD MARK (glycols, oxoproline, L-lactate, D-lactate, methanol, aspirin, renal failure, ketoacidosis). A normal anion-gap metabolic acidosis (a hyperchloraemic acidosis) means bicarbonate has been lost or chloride gained: the fall in bicarbonate is matched by a rise in chloride, so the gap is unchanged. The causes are diarrhoea, renal tubular acidosis, acetazolamide, ureteric diversion, and the chloride loading of normal saline resuscitation [1][10].

Two corrections are essential. First, hypoalbuminaemia lowers the apparent anion gap: each 10 g per litre fall in albumin lowers the normal gap by about 2.5 mmol per litre, so the corrected gap is the observed gap plus 2.5 times (4 minus albumin in g per decilitre). Failing to correct in a hypoalbuminaemic intensive-care patient will miss a high-gap acidosis. Second, calculation of the delta ratio (above) lets a high-gap and a normal-gap acidosis coexist in the same patient, which is common [6][10].

The Davenport diagram

The Davenport diagram is the geometric device that ties the whole topic together, and a candidate who can draw it has effectively answered the topic. The axes are pH on the x (from about 7.0 to 7.8) and bicarbonate on the y (from 0 to about 50 mmol per litre). On it are drawn two families of lines [1][5].

The buffer line is a near-straight diagonal with a negative slope that passes through the normal point (pH 7.40, bicarbonate 24). Its slope is set by the non-bicarbonate buffers (chiefly haemoglobin): the more non-bicarbonate buffer, the steeper the line. It represents the locus of points reachable by changing PaCO2 alone at a fixed buffer base, which is precisely what a respiratory disorder does. A respiratory acidosis (high PaCO2) moves the point up the buffer line to the left (lower pH, higher bicarbonate); a respiratory alkalosis (low PaCO2) moves it down the buffer line to the right (higher pH, lower bicarbonate). Movement along the line is the signature of a pure respiratory process. [1]

The bicarbonate isobar (or PCO2 isobar) curves describe the bicarbonate at each pH for a fixed PaCO2; the PaCO2 40 isobar passes through the normal point. A metabolic disorder, which changes the bicarbonate at a constant PaCO2, moves the point off the buffer line: a metabolic acidosis shifts the whole buffer line downward (parallel, lower bicarbonate at every pH), and a metabolic alkalosis shifts it upward (parallel, higher bicarbonate at every pH). Compensation then draws the point back towards the normal region along the new buffer line: renal compensation of a respiratory acidosis moves the point down-right towards normal, and respiratory compensation of a metabolic acidosis moves it down-left along a lower buffer line. The compensation bands on the diagram are the confidence regions for each disorder with appropriate compensation, and a point lying outside all bands means a mixed disorder [1][5].

[1]

Base excess and the systematic ABG method

Every arterial blood gas can be read by a fixed five-step method, and applying it consistently is the difference between a diagnosis and a guess [2][4].

The method is as follows. Step one, read the pH. A pH below 7.35 is an acidaemia; above 7.45 an alkalaemia; if normal with an abnormal PaCO2 and bicarbonate, the disorder is mixed. Step two, read the PaCO2 for the respiratory component: it is high in respiratory acidosis and low in respiratory alkalosis. Step three, read the bicarbonate or base excess for the metabolic component: a low bicarbonate or negative base excess is a metabolic acidosis; a high bicarbonate or positive base excess is a metabolic alkalosis. Match the direction of the pH change to the apparent primary disorder: the process that pulls the pH in the observed direction is the primary one [2][4].

Step four, calculate the expected compensation. For a metabolic acidosis apply Winter's formula; for a metabolic alkalosis apply the 0.7 rule or the 55 mmHg ceiling; for a respiratory acidosis or alkalosis decide whether the process is acute or chronic and apply the per-10-mmHg rule. Compare the measured value with the expected: concordance means appropriate compensation, discordance means a mixed disorder. Step five, calculate the anion gap (and correct for albumin), then the delta ratio if there is a high-gap acidosis. This step turns a single metabolic acidosis into a cause, and unmasks the mixed metabolic disorders [1][2].

Worked examples

Four worked examples illustrate the method and are worth committing to memory as archetypes [1][2].

Example one, the diabetic ketoacidosis. pH 7.20, PaCO2 20 mmHg, bicarbonate 8 mmol per litre, base excess minus 20. Step one: severe acidaemia. Step two: PaCO2 is low, so the respiratory component would cause an alkalosis, opposite to the pH, so the primary disorder is metabolic. Step three: bicarbonate and base excess are very low, confirming a metabolic acidosis. Step four: Winter's formula gives expected PaCO2 of 1.5 times 8 plus 8, which is 20, plus or minus 2; the measured 20 is within the band, so compensation is appropriate. Step five: the anion gap is high (ketoacidosis). This is a pure high-anion-gap metabolic acidosis with appropriate Kussmaul compensation [1].

Example two, the acute opioid overdose. pH 7.20, PaCO2 80 mmHg, bicarbonate 29 mmol per litre, base excess plus 1. Step one: acidaemia. Step two: PaCO2 is very high, in the direction of the pH, so the primary disorder is a respiratory acidosis. Step three: bicarbonate is only slightly raised and base excess is normal, so there is no metabolic component. Step four: for an acute respiratory acidosis, the PaCO2 has risen by 40 mmHg above normal (80 minus 40), so the expected bicarbonate rise is about 4 mmol per litre (1 per 10), giving an expected bicarbonate of about 28; the measured 29 fits. Step five: the anion gap is normal. This is an acute uncompensated respiratory acidosis [2].

Example three, the chronic COPD patient. pH 7.36, PaCO2 65 mmHg, bicarbonate 36 mmol per litre, base excess plus 10. Step one: mild acidaemia (near-normal pH). Step two: PaCO2 high, in the direction of the pH, so the primary disorder is a respiratory acidosis. Step three: bicarbonate and base excess are raised, indicating renal compensation. Step four: the PaCO2 has risen by 25 mmHg above normal; chronic compensation predicts a bicarbonate rise of about 10 mmol per litre (4 per 10), giving an expected bicarbonate of about 34, close to the measured 36. Step five: anion gap normal. This is a chronic respiratory acidosis with appropriate renal compensation. The contrast with example two is the point: the same PaCO2 gives a much lower pH acutely than chronically [8].

Example four, the septic surgical patient. pH 7.30, PaCO2 28 mmHg, bicarbonate 14 mmol per litre, base excess minus 12, lactate 5 mmol per litre. Step one: acidaemia. Step two: PaCO2 low, opposite to the pH, so the primary disorder is metabolic. Step three: bicarbonate and base excess low, metabolic acidosis. Step four: Winter's formula gives expected PaCO2 of 1.5 times 14 plus 8, which is 29, plus or minus 2; the measured 28 is within the band. Step five: the anion gap is high (lactate). This is a high-anion-gap lactic acidosis from sepsis with appropriate respiratory compensation [10].

The Stewart approach (brief cross-link)

An alternative analysis, the Stewart or strong-ion approach, reframes acid-base in terms of three independent variables: the strong-ion difference (chiefly sodium minus chloride), the total weak acid (albumin and phosphate), and the PaCO2. It resolves several ambiguities of the bicarbonate method, notably the effect of hypoalbuminaemia (which causes a metabolic alkalosis in the Stewart view that is hidden in the conventional approach) and of a chloride-rich fluid like normal saline (which causes a hyperchloraemic acidosis by narrowing the strong-ion difference). The two frameworks agree on the four primary disorders and their compensation; Stewart adds mechanistic depth around albumin and chloride and is covered in its own topic. The practical rule is that a hyperchloraemic acidosis after saline resuscitation is real and is best understood as a reduction in the strong-ion difference, not as simple bicarbonate dilution [5][6].

Anaesthetic and perioperative relevance

• Respiratory acidosis is the commonest perioperative disorder: hypercapnia from hypoventilation under opioids, residual neuromuscular blockade, deep anaesthesia, or inadequate ventilator settings. The pH falls briskly because the kidney has not yet compensated, and the rise in PaCO2 drives a rise in cerebral blood flow and intracranial pressure, which matters in head injury.
• Metabolic acidosis with a raised anion gap and lactate signals tissue hypoperfusion (shock, sepsis, ischaemia, mesenteric or limb, prolonged tourniquet) and is a marker to resuscitate and find the cause, not merely to treat a number. A widening base deficit during surgery predicts morbidity and mortality.
• Metabolic alkalosis from gastric drainage or vomiting is common postoperatively and corrects with volume and chloride (the saline-responsive alkalosis); a persisting alkalosis suggests ongoing diuretic use or mineralocorticoid excess.
• Hyperchloraemic metabolic acidosis follows large-volume normal saline resuscitation (the strong-ion difference narrows); balanced crystalloids (Plasma-Lyte, Hartmann) reduce its incidence and are preferred when more than a litre or two is given.
• Mixed disorders are the rule in the critically ill: the septic patient with a lactic acidosis and a stress-induced respiratory alkalosis; the COPD patient on diuretics with a respiratory acidosis and a metabolic alkalosis; the salicylate-overdose patient with a high-gap acidosis and a respiratory alkalosis. Always apply the compensation rules and the delta ratio.
• The renal compensation rules are timed: a metabolic disorder compensates within hours, a respiratory disorder over days. Quoting the chronic compensation rule for an acute respiratory disorder is a frequent and serious error.
• Chronic buffering uses bone: a patient with chronic metabolic acidosis (chronic kidney disease) buffers partly by bone carbonate dissolution, contributing to osteopaenia and hypercalciuria [1][2][10].

The synthesis

Acid-base homeostasis reduces to a ratio and two organs. The ratio is bicarbonate to dissolved CO2, set by the Henderson-Hasselbalch equation, and the two organs are the lung (fast, excreting CO2) and the kidney (slow, regenerating bicarbonate). The buffers distribute the acid load, with the open bicarbonate system dominating the extracellular fluid, haemoglobin the red cell (and the Haldane effect), phosphate the cell and the urine, and bone the chronic reservoir. The four primary disorders are read off the pH, PaCO2 and bicarbonate; the expected compensation is read off a small set of rules; and base excess and the anion gap refine the metabolic component. Systematic application of the five-step method converts any blood gas into a named disorder with appropriate, deficient, or mixed compensation [1][3][4].

Red flags

• pH is set by the RATIO of bicarbonate to (0.03 times PaCO2), not by either alone; a normal pH with an abnormal PaCO2 and bicarbonate is always mixed, never fully compensated.
• Respiratory compensation is fast (minutes); renal compensation is slow (two to five days), so an acute respiratory disorder is poorly compensated and a chronic one well compensated.
• Winter's formula for metabolic acidosis: expected PaCO2 equals 1.5 times bicarbonate plus 8, plus or minus 2 mmHg. Always check it; a PaCO2 outside the band means a mixed disorder.
• Respiratory acidosis: bicarbonate up about 1 acutely and about 4 chronically per 10 mmHg rise in PaCO2. The same PaCO2 gives a much lower pH acutely than chronically.
• Base excess is the metabolic component at a standardised PaCO2 of 40 mmHg, independent of the respiratory picture; a negative base excess means a metabolic acidosis regardless of PaCO2.
• The anion gap separates metabolic acidoses into high-gap (added acid, MUDPILES) and normal-gap (bicarbonate loss or chloride gain). Correct for albumin: the gap falls about 2.5 mmol per litre for each 10 g per litre fall in albumin.
• Bicarbonate buffers poorly at pH 7.4 by pKa but dominates because it is open: the lung exhales the CO2 and the kidney regenerates the bicarbonate. A closed bicarbonate buffer would be near-useless.
• Deoxyhaemoglobin is a better buffer than oxyhaemoglobin (the Haldane effect), which is why venous blood carries more CO2 than arterial at the same PCO2. [1]

References and further reading

1. Adrogué HJ, Madias NE. Management of life-threatening acid-base disorders. First of two parts. N Engl J Med 1998;338(1):26-34. PMID 9414329.
2. Adrogué HJ, Madias NE. Management of life-threatening acid-base disorders. Second of two parts. N Engl J Med 1998;338(2):107-111. PMID 9420343.
3. Berend K. Diagnostic use of base excess in acid-base disorders. N Engl J Med 2018;378(15):1419-1428. PMID 29641969.
4. Story DA. Bench-to-bedside review: a brief history of clinical acid-base. Crit Care 2004;8(4):253-258. PMID 15312207.
5. Kellum JA. Determinants of plasma acid-base balance. Crit Care Clin 2005;21(2):329-346. PMID 15781166.
6. Figge J, Rossing TH, Fencl V. The role of serum proteins in acid-base equilibria. J Lab Clin Med 1991;117(6):453-467. PMID 2045713.
7. Madias NE, Adrogué HJ. Cross-talk between two organs: how the kidney responds to disruption of acid-base balance by the lung. Nephron Physiol 2003;93(3):p61-66. PMID 12660492.
8. Madias NE. Renal acidification responses to respiratory acid-base disorders. J Nephrol 2010;23 Suppl 16:S85-91. PMID 21170892.
9. Fulop M. A guide for predicting arterial CO2 tension in metabolic acidosis. Am J Nephrol 1997;17(5):421-424. PMID 9382159.
10. Kraut JA, Madias NE. Lactic acidosis: current treatments and future directions. Am J Kidney Dis 2016;68(3):473-482. PMID 27291485.
11. Lang W, Zander R. The accuracy of calculated base excess in blood. Clin Chem Lab Med 2002;40(4):404-410. PMID 12059083.
12. Berend K. Acid-base pathophysiology after 130 years: confusing, irrational and controversial. J Nephrol 2013;26(2):254-265. PMID 22976522. [1]

References

1. [1]Adrogué HJ, Madias NE Management of life-threatening acid-base disorders. First of two parts N Engl J Med, 1998.PMID 9414329
2. [2]Adrogué HJ, Madias NE Management of life-threatening acid-base disorders. Second of two parts N Engl J Med, 1998.PMID 9420343
3. [3]Berend K Diagnostic Use of Base Excess in Acid-Base Disorders N Engl J Med, 2018.PMID 29641969
4. [4]Story DA Bench-to-bedside review: a brief history of clinical acid-base Crit Care, 2004.PMID 15312207
5. [5]Kellum JA Determinants of plasma acid-base balance Crit Care Clin, 2005.PMID 15781166
6. [6]Figge J, Rossing TH, Fencl V The role of serum proteins in acid-base equilibria J Lab Clin Med, 1991.PMID 2045713
7. [7]Madias NE, Adrogué HJ Cross-talk between two organs: how the kidney responds to disruption of acid-base balance by the lung Nephron Physiol, 2003.PMID 12660492
8. [8]Madias NE Renal acidification responses to respiratory acid-base disorders J Nephrol, 2010.PMID 21170892
9. [9]Fulop M A guide for predicting arterial CO2 tension in metabolic acidosis Am J Nephrol, 1997.PMID 9382159
10. [10]Kraut JA, Madias NE Lactic Acidosis: Current Treatments and Future Directions Am J Kidney Dis, 2016.PMID 27291485
11. [11]Lang W, Zander R The accuracy of calculated base excess in blood Clin Chem Lab Med, 2002.PMID 12059083
12. [12]Berend K Acid-base pathophysiology after 130 years: confusing, irrational and controversial J Nephrol, 2013.PMID 22976522