Respiratory Acid–Base Disorders and Compensatory Responses

A change in carbon dioxide tension can alter blood pH rapidly, while renal and whole-body adaptation evolve along a different trajectory. Interpreting a respiratory acid–base disorder therefore requires both the current chemical state and the sequence that produced it. This chapter connects effective ventilation, buffering, renal acid balance, and organ effects before examining the limits of expected compensation. In the constructed gas examples, bicarbonate denotes gas-calculated bicarbonate rather than separately measured chemistry-panel total carbon dioxide; the selected pairs are checked with the stated chemical coefficients.

56.1 Carbon dioxide tension expresses a production–elimination relationship

A respiratory acid–base process begins when carbon dioxide availability changes relative to its elimination. Hypercapnia means increased arterial carbon dioxide tension; hypocapnia means decreased tension. Neither term identifies the mechanism. Respiratory muscle failure, depressed respiratory drive, increased apparatus dead space, severe ventilation–perfusion inequality, and increased metabolic production can all change the balance. A patient may breathe rapidly yet eliminate carbon dioxide inadequately because much of each breath ventilates conducting airways or poorly perfused alveoli.

At a steady state, metabolic carbon dioxide production equals net pulmonary elimination if there is no important extracorporeal route. The familiar alveolar ventilation relationship describes this balance. Let PaCO₂ be arterial carbon dioxide tension in mmHg, V̇CO₂ carbon dioxide elimination in mL/min expressed at standard temperature and pressure, dry, and V̇A effective alveolar ventilation in L/min expressed at body temperature and pressure, saturated. With negligible inspired carbon dioxide, the unit-conversion coefficient is 0.863:

For an illustrative elimination of 200 mL/min and effective alveolar ventilation of 4 L/min, PaCO₂ is approximately 0.863 × 200 / 4 = 43.2 mmHg. Reducing effective ventilation to 3 L/min gives 57.5 mmHg. Alternatively, increasing elimination to 300 mL/min while ventilation remains 4 L/min gives 64.7 mmHg. Thus a rise in carbon dioxide tension can reflect less elimination capacity, more metabolic burden, or both. These constructed values illustrate proportionality rather than normal limits or ventilator prescriptions.

The equation requires care during transitions. Carbon dioxide can enter or leave large bicarbonate and tissue stores, so exhaled elimination temporarily differs from metabolic production. Restoration of blood flow may wash accumulated carbon dioxide from tissues without a matching increase in current cellular production. Inspired carbon dioxide or rebreathing adds another source. Moreover, an injured lung is not one uniform alveolus: the “effective” ventilation in this relationship summarizes a heterogeneous exchange system. Total minute ventilation alone therefore cannot explain either PaCO₂ or the effort required to achieve it.

56.2 A respiratory process and the resulting pH are different observations

Increasing dissolved carbon dioxide shifts the bicarbonate reaction toward hydrogen ions and bicarbonate. Decreasing carbon dioxide shifts it in the opposite direction. Respiratory acidosis and respiratory alkalosis name these acidifying and alkalinizing processes. Acidemia and alkalemia describe the measured blood pH. A respiratory acidosis can coexist with an alkaline pH when another alkalinizing process is stronger; a low PaCO₂ can accompany acidemia when ventilation is responding to metabolic acid accumulation.

The bicarbonate relationship makes the distinction explicit. Here pH is the logarithmic hydrogen-ion activity measure, [HCO₃⁻] is plasma bicarbonate concentration in mmol/L, and PaCO₂ is in mmHg. At approximately 37°C, use a rounded apparent pK of 6.1 and carbon dioxide solubility coefficient of 0.030 mmol/L/mmHg. The ratio inside the logarithm is dimensionless:

An illustrative bicarbonate concentration of 24 mmol/L with PaCO₂ of 40 mmHg gives 6.1 + log₁₀[24/(0.030 × 40)] = 7.40. The same PaCO₂ paired with a different bicarbonate concentration gives a different pH. Conversely, similar pH values can conceal very different carbon dioxide tensions and bicarbonate pools. The relationship describes the chemical state; it does not establish which physiological disturbance occurred first.

This matters when interpreting ventilatory responses. A low PaCO₂ during metabolic acidosis may represent an appropriate increase in effective ventilation, an additional respiratory alkalinizing process, or both. Determining which interpretation fits requires the metabolic disturbance, prior measurements, time course, and expected response. Calling every low PaCO₂ “primary respiratory alkalosis” confuses a measured gas tension with causal sequence.

The numerical coefficients are approximations for stated temperature and conventions. They should not be mixed indiscriminately with temperature-corrected gas values or bicarbonate calculated using a different laboratory model. More fundamentally, arterial pH is a sample of one compartment. It provides essential information without measuring the hydrogen-ion activity inside every organ or quantifying the amount of retained acid in the body.

56.3 The immediate hypercapnic response is predominantly buffering

When PaCO₂ rises rapidly, carbon dioxide crosses cell membranes and hydrates within blood and tissues. Carbonic anhydrase accelerates equilibration in erythrocytes and other sites. Hydrogen ions bind to nonbicarbonate buffers, including hemoglobin and intracellular proteins. Bicarbonate therefore increases even before the kidneys have added substantial new base to the body. This early bicarbonate increase is evidence of chemical redistribution and buffering, not proof of renal compensation.

Erythrocytes participate through coupled processes. Bicarbonate movement and chloride exchange distribute the response between cells and plasma. Hemoglobin buffering depends partly on oxygenation; oxygen loading and unloading alter carbon dioxide and hydrogen-ion carriage through the Haldane effect. These processes occur alongside changes in tissue carbon dioxide storage. Consequently, neither plasma bicarbonate nor total blood carbon dioxide behaves as an isolated, fixed container.

Elkinton and colleagues studied acute experimental hypercapnia and hypocapnia in healthy men, assessing ionic transfers across body-fluid compartments. Their observations supported participation of tissue buffering beyond the extracellular bicarbonate system. The experimental preparation, including bicarbonate loading in several hypercapnic studies, matters when interpreting the findings; it was not an unselected cohort of critically ill patients. 1

A conventional teaching approximation is an increase in plasma bicarbonate of about 1 mmol/L for each 10 mmHg acute increase in PaCO₂. Starting from a constructed baseline of 40 mmHg and 24 mmol/L, an increase to 60 mmHg would therefore be paired illustratively with bicarbonate of 26 mmol/L. The pH calculation is 6.1 + log₁₀[26/(0.030 × 60)] = 7.26. Buffering has moderated the fall without restoring the original pH.

This approximation is a comparison aid rather than an exact experimentally determined response for every patient. Initial buffer concentrations, hemoglobin, temperature, concurrent metabolic changes, and the speed and magnitude of the carbon dioxide change influence the result. In particular, a patient beginning with chronic hypercapnia should be compared with that patient's established state. Returning automatically to an assumed 40/24 baseline can manufacture a metabolic diagnosis that the observed trajectory does not support.

56.4 Renal adaptation changes bicarbonate balance and acid excretion

Sustained hypercapnia recruits renal processes that differ from immediate buffering. Filtered bicarbonate must be reclaimed to prevent urinary base loss. Additional bicarbonate can enter the body when net renal acid excretion increases, including ammonium excretion and hydrogen-ion excretion with urinary buffers. Reclaiming an existing filtered bicarbonate ion and generating net new bicarbonate are therefore distinct contributions. Chapter 52 develops the accounting in detail.

Carbon dioxide and intracellular acid–base changes influence tubular transport, while potassium balance, chloride availability, hormonal signals, and renal perfusion modify the response. Increased proximal ammonium production and coordinated medullary recycling and collecting-duct transport help support acid elimination. Changes in bicarbonate secretion can also alter net base retention. The renal response is not simply a higher glomerular filtration rate or a uniform increase in every component of urinary acid.

The temporal division between “immediate” and “renal” responses is useful but imperfect. Barker and colleagues found renal changes during short experimental respiratory disturbances in humans, including altered urinary bicarbonate and acid excretion. Their preparation overlapped the carefully controlled, sometimes preloaded experiments described above. It demonstrates that a renal response can begin promptly; it does not define a universal interval before compensation starts. 2

Original sustained-hypercapnia experiments in dogs showed an evolving combination of buffering, electrolyte changes, and renal acid–base adjustment. 3 More recent mouse experiments found increased ammonium excretion and changes in transport proteins during hypercapnia, with unchanged titratable acid under those conditions. These findings support a coordinated renal mechanism while demonstrating why species, exposure, and the particular urinary component must be specified. 4

During establishment of adaptation, extra net acid excretion can increase the body's bicarbonate reserve. Once a new steady state is reached, sustained excess excretion cannot continue indefinitely without further changing body composition. Excretion again has to match ongoing acid production and intake, but around a different bicarbonate concentration and carbon dioxide tension. Thus a single urine sample cannot reconstruct the entire period over which adaptation developed.

56.5 Sustained hypercapnia has an empirical, population-dependent response

The usual distinction between acute and chronic respiratory acidosis describes the degree of adaptation as well as duration. Sustained carbon dioxide elevation generally permits a larger bicarbonate increase and a smaller pH fall than the immediate response. “Chronic” should nevertheless be supported by history and serial measurements. A high bicarbonate concentration alone could also reflect chloride depletion, potassium depletion, administered base, or another alkalinizing process.

Conventional teaching often estimates a bicarbonate increase of approximately 3.5–4 mmol/L per 10 mmHg sustained increase in PaCO₂. A prospective human study illustrates why this cannot be treated as an invariant biological constant. González and colleagues studied 28 selected, stable outpatients with chronic hypercapnia, preserved kidney function, and exclusions intended to minimize other acid–base influences. Their fitted bicarbonate slope was approximately 0.476 mmol/L per mmHg; the authors proposed a rounded 0.5 relationship within the range studied. Their observations did not extend beyond PaCO₂ of 70 mmHg. 5

Consider the consequences for an illustrative comparison beginning at bicarbonate 24 mmol/L and PaCO₂ 40 mmHg. At PaCO₂ 60 mmHg, a slope of 0.4 gives bicarbonate 24 + 0.4 × 20 = 32 mmol/L; a rounded slope of 0.5 gives 34 mmol/L. The corresponding calculated pH values are approximately 7.35 and 7.38. Both responses differ substantially from the acute illustrative pair of 60 mmHg and 26 mmol/L.

These calculations compare possible response relationships; they are not competing treatment targets. The actual patient may differ in kidney function, dietary acid load, medications, oxygenation, or the stability of carbon dioxide exposure. A fitted mean and its prediction interval describe a particular sample under specified conditions. They do not prove that every value inside the interval represents one uncomplicated process or that every value outside it requires a second diagnosis.

Extrapolation is especially hazardous in extreme hypercapnia or rapidly changing critical illness. The relevant questions are whether the gas tension was sustained, whether the kidney could respond, and whether the measured bicarbonate trajectory is coherent with the patient's other balances.

56.6 Acute hypocapnia draws on the same buffers in the opposite direction

A rapid fall in PaCO₂ reduces dissolved carbon dioxide and shifts the bicarbonate reaction toward carbon dioxide formation. Nonbicarbonate buffers release hydrogen ions, consuming some bicarbonate and limiting the rise in pH. Intracellular and extracellular buffering again interact. The immediate decrease in plasma bicarbonate therefore does not by itself indicate loss of bicarbonate from the body or a newly generated metabolic acidosis.

A conventional acute comparison uses a bicarbonate decrease of approximately 2 mmol/L for each 10 mmHg decrease in PaCO₂. From the illustrative starting pair of 40 mmHg and 24 mmol/L, a fall to 30 mmHg gives bicarbonate of approximately 22 mmol/L. Using the stated 37°C coefficients, pH is 6.1 + log₁₀[22/(0.030 × 30)] = 7.49. The bicarbonate fall partly buffers the alkalinizing carbon dioxide change, but the pH still rises.

The asymmetry between the familiar acute hypercapnic and hypocapnic numerical aids should discourage thinking of compensation as one linear rule over the entire carbon dioxide range. Buffer titration is not perfectly linear, and the starting acid–base state matters. These rounded aids are useful near their customary teaching conditions, not as predictions of the pH attained by any specified change in ventilation.

The mechanism producing hypocapnia also remains important. Hypoxemia, central nervous system disease, pain, inflammatory illness, pregnancy-related ventilatory changes, and externally assisted ventilation can all increase ventilation relative to carbon dioxide production. The label “hyperventilation” identifies that relative excess; it does not establish a psychological cause. A reduction in metabolic production or a change in circulation and carbon dioxide delivery can alter the observed gas tension as well.

Moreover, hypocapnia may be serving a compensatory role in metabolic acidosis. Its presence should prompt assessment of why ventilation increased and whether the increase is proportionate to the metabolic disturbance. An apparently modest PaCO₂ can represent inadequate elimination for that context, while an unusually low value may reflect an additional respiratory stimulus. Those comparisons require the broader interpretation developed in Chapter 57.

56.7 Sustained hypocapnia permits renal reduction of the bicarbonate reserve

With sustained hypocapnia, renal adaptation moves in the opposite direction from the hypercapnic response. Reduced hydrogen-ion secretion and altered bicarbonate handling lower net acid excretion and can permit urinary bicarbonate loss during establishment of the new state. Body bicarbonate availability falls, moderating the alkalemia. As with hypercapnia, the eventual maintenance state should not be confused with the transient imbalance that established it.

Krapf and colleagues studied sustained respiratory alkalosis in nine healthy young men exposed to altitude-related hypobaric hypoxia; five were also studied with experimentally induced metabolic acidosis. Plasma bicarbonate fell by about 0.41 mmol/L per mmHg fall in PaCO₂ without the imposed acidosis and by a similar slope with it. The study directly linked sustained human hypocapnia to a substantial bicarbonate response, while retaining the important limitations of a small, selected, altitude-exposed population. 6

For an illustrative sustained-response calculation, use a rounded decrease of 4 mmol/L per 10 mmHg decrease in PaCO₂. Starting at 40 mmHg and 24 mmol/L, a PaCO₂ of 30 mmHg is paired with bicarbonate of 20 mmol/L. The calculated pH is 6.1 + log₁₀[20/(0.030 × 30)] = 7.45, compared with approximately 7.49 in the acute illustrative case. The adaptation reduces the pH displacement without requiring exact restoration of its initial value.

This numerical relationship should not be assumed identical in sepsis, pregnancy, liver disease, mechanical ventilation, and high-altitude exposure. Each setting can alter acid production, kidney function, weak acids, or strong-ion composition. The finding of a lower bicarbonate concentration in a hypocapnic patient must therefore be interpreted against both the probable duration and these concurrent influences.

Recovery also has its own trajectory. If carbon dioxide tension returns toward its earlier value before the reduced bicarbonate reserve is restored, the patient can become acidemic. That possibility follows from unequal recovery of the respiratory and metabolic components, not from failure of the bicarbonate equation. A single label such as “resolved respiratory alkalosis” may obscure an adaptation that remains physiologically active.

56.8 Transitions must be interpreted from the patient's previous state

Acute changes frequently occur on top of established adaptation. Consider a patient with documented stable PaCO₂ of 60 mmHg and bicarbonate of 32 mmol/L. If PaCO₂ rapidly rises to 80 mmHg, the acute teaching increment of 1 mmol/L bicarbonate per 10 mmHg suggests an illustrative bicarbonate of 34 mmol/L. The calculated pH is 6.1 + log₁₀[34/(0.030 × 80)] = 7.25. This is acute respiratory acidification superimposed on an adapted baseline, not a newly established chronic state at 80 mmHg.

The example uses the actual prior bicarbonate concentration. Estimating chronic adaptation anew from 40 mmHg would disregard the transition and might incorrectly attribute the difference to an additional metabolic acidosis. Equally, a measured result that differs from 34 mmol/L requires consideration of real buffering variability, concurrent metabolic events, and sampling before a second process is declared.

The reverse transition can uncover residual base. Suppose a later arterial sample has PaCO₂ of 40 mmHg and gas-calculated bicarbonate of 30 mmol/L. Its pH is 6.1 + log₁₀[30/(0.030 × 40)] = 7.50. The example deliberately specifies the later bicarbonate rather than assuming that the previous 32 mmol/L stays fixed: rapid nonbicarbonate buffering also reverses when carbon dioxide falls. Renal base adjustment, chloride and potassium balance, and changes in extracellular volume can then influence the subsequent course.

Original recovery experiments in chronically hypercapnic dogs demonstrated that chloride intake modified the persistence and correction of elevated bicarbonate after carbon dioxide exposure ended. They support a physiological interaction between prior adaptation and electrolyte balance, while providing no human correction schedule or fluid prescription. 7 Chapter 55 develops the renal mechanisms of posthypercapnic metabolic alkalosis.

Similar reasoning applies after sustained hypocapnia: the respiratory change can disappear before the associated bicarbonate adjustment. Mechanical assistance, sedation changes, altered respiratory workload, and recovery from disease can move PaCO₂ faster than whole-body electrolyte and acid balances recover. Interpretation should follow this sequence rather than forcing every sample into an isolated acute-or-chronic category.

56.9 Expected compensation is a comparison distribution, not a diagnostic verdict

An expected-response rule asks how much a secondary variable commonly changes when a specified primary disturbance occurs under selected conditions. It does not demonstrate that the primary disturbance is the only process present. Published human response analyses depend on patient selection, measured variables, and the definition of steady state; an analysis using standard base excess, for example, is not numerically interchangeable with a plasma-bicarbonate slope. 8

The arithmetic can be written generically. Let Δ[HCO₃⁻] be the change in plasma bicarbonate concentration in mmol/L, ΔPaCO₂ the change in arterial carbon dioxide tension in mmHg, and s an empirical response slope in mmol/L/mmHg for the stated setting:

This form makes the required assumptions visible. A baseline must be specified, the slope must match the direction and degree of adaptation, and both measurements must represent comparable conditions. The rounded acute hypercapnic slope used above is 0.1; the acute hypocapnic comparison is 0.2, with a negative carbon dioxide change producing a negative bicarbonate change. The sustained slopes are larger, but their numerical values vary across the evidence.

Several limitations then follow. A prediction band describes dispersion among individuals, not an acceptable pH range or a treatment target. Error in pH and PaCO₂ also affects calculated bicarbonate, creating mathematical dependence between variables. Prior base loss, chloride depletion, kidney dysfunction, and changes in albumin may shift the starting point without obeying the chosen slope. During a transition, neither an acute endpoint nor a sustained endpoint may yet apply.

Opposing metabolic disturbances can also cancel numerically. A bicarbonate value inside an expected band therefore cannot exclude two additional processes. A value outside it is a reason to inspect the clinical trajectory and other measurements, rather than an automatic proof of one specific metabolic disorder. The appropriate output is a physiologically supported hypothesis that can be tested against history, electrolyte balance, and serial results.

56.10 Arterial carbon dioxide and organ acid–base conditions can diverge

Carbon dioxide crosses biological membranes readily, but tissue pH also depends on local buffers, ion transport, metabolism, and perfusion. The response of brain, muscle, and blood therefore need not be identical. An arterial gas describes blood after pulmonary exchange, whereas an organ is exposed to its own balance of carbon dioxide production and removal. During changing flow, the distinction becomes especially important.

In a small human experiment, Cadoux-Hudson and colleagues used phosphorus magnetic resonance spectroscopy during an inhaled hypercapnic load and observed brain acidification, with regional differences. The controlled exposure in healthy volunteers supports a compartmental response; it does not establish a tissue-pH prediction for critically ill patients. 9 Classic direct human measurements also demonstrated cerebral blood-flow responses to altered carbon dioxide tension. 10 Thus a carbon dioxide change modifies both local chemistry and its vascular environment.

Systemic circulation adds another layer. With low tissue blood flow, carbon dioxide removal can be inadequate even when ventilated arterial blood has a relatively low PaCO₂. Adrogué and colleagues obtained simultaneous arterial and central venous samples across different circulatory states and documented important acid–base disparities in severe circulatory failure. 11 The venous abnormality cannot be translated directly into the pH of every cell, but the observation shows why an arterial result can incompletely represent systemic carbon dioxide stagnation.

Oxygenation, pulmonary vascular load, respiratory effort, and metabolic demand interact with these changes. Increased effort can raise carbon dioxide production and oxygen consumption. Hypoxemia may stimulate ventilation while also altering metabolism and hemoglobin buffering. Hypercapnic acidification can influence myocardial performance and vascular tone, but the net response depends on sympathetic activation, existing cardiac disease, and the pulmonary circulation.

Consequently, the clinical significance of a PaCO₂ cannot be read from its number alone. Its speed of change, associated pH, cerebral and cardiovascular context, oxygen delivery, and capacity for compensation all matter. Physiological effects establish reasons for assessment; they do not establish a universal harmless or dangerous carbon dioxide threshold.

56.11 Measurement site and timing belong inside the interpretation

Blood-gas interpretation begins with knowing what was measured, where it was sampled, and when. Arterial PaCO₂ characterizes the respiratory component after lung exchange. Peripheral or central venous carbon dioxide tension additionally reflects local or systemic carbon dioxide delivery and blood flow. These samples can be useful, but they are not interchangeable by applying one fixed subtraction to every patient.

A prospective study of 234 patients with COPD exacerbations found useful arterial–venous agreement for pH and bicarbonate in its assessment setting. That result concerns specified variables in a particular population; it does not establish universal exchangeability of all venous and arterial gas measurements. 12 Circulatory failure further changes the relation, as the paired-sampling observations discussed above illustrate. The clinical question determines which compartment must be measured.

End-tidal carbon dioxide supplies a valuable breath-by-breath signal, but it represents gas at the airway late in expiration. Changes in alveolar dead space, uneven emptying, tidal volume, leaks, and pulmonary perfusion can change its relationship with PaCO₂. A sudden end-tidal fall may reflect altered blood flow rather than a proportional fall in total body carbon dioxide. It should not automatically be entered into an arterial compensation equation.

Blood-gas bicarbonate is usually calculated from measured pH and PCO₂. It is therefore not a third independent confirmation of those two measurements. Chemistry-panel total carbon dioxide is separately measured and includes dissolved carbon dioxide as well as bicarbonate. Different sampling times, sample exposure to air, delays, and handling can produce discrepancies. Chapter 53 explains the chemical and analytical distinctions; here their practical implication is to compare contemporaneous, properly identified measurements before assigning a physiological cause.

Timing is particularly consequential during changing respiratory support, effort, circulation, or metabolic acid production. A gas drawn before a major change and electrolytes drawn afterward may form a chemically plausible but physiologically nonexistent composite. Temperature conventions should also be identified consistently. Serial samples become informative when linked to events, because the observed direction and speed of change often discriminate among mechanisms more effectively than a single apparently precise compensation calculation.

56.12 Impaired adaptation and concurrent processes require an integrated explanation

Kidney dysfunction can limit bicarbonate conservation, ammonium production and excretion, and removal of nonvolatile acid. A patient with respiratory acidosis and renal failure may therefore show less bicarbonate adaptation than an otherwise comparable patient with preserved kidney function. The deficit can combine limited secondary response with a distinct metabolic acidifying process. These mechanisms overlap clinically and cannot always be partitioned by one formula.

Consider a patient with documented stable PaCO₂ of 60 mmHg and gas-calculated bicarbonate of 32 mmol/L. Later, PaCO₂ remains 60 mmHg but gas-calculated bicarbonate is 22 mmol/L. The earlier calculated pH was 7.35; the later value is 6.1 + log₁₀[22/(0.030 × 60)] = 7.19. The new acidemia cannot be explained by a further rise in arterial carbon dioxide because none occurred. A change in bicarbonate balance or other nonrespiratory influences must be investigated.

That conclusion does not identify its cause. Retained nonvolatile acids, gastrointestinal base loss, altered chloride balance, a change in administered solutions, organic anion accumulation, and changing protein concentrations require different evidence. The clinician should examine renal function, electrolytes, albumin, lactate and other relevant measurements, treatment exposures, and the time course. Likewise, unexpectedly high bicarbonate during hypercapnia may reflect an additional alkalinizing process or persisting adaptation from an earlier, higher PaCO₂.

A coherent interpretation therefore combines three questions: what changed carbon dioxide production or effective elimination, how much rapid buffering and renal adaptation was possible, and what other balances changed concurrently? Answering them may explain the trajectory without assuming that every discrepancy is a new primary disease. Conversely, apparently expected compensation should not suppress evidence of an independent metabolic process.

The purpose of this analysis is to describe mechanism and reserve. It does not specify a ventilatory correction rate, a desired PaCO₂, or a bicarbonate target. Those decisions require the underlying disease, organ effects, competing risks, and relevant clinical evidence. Chapter 57 extends this framework to the formal interpretation of mixed disorders, using the same requirement that numerical patterns be reconciled with physiology and serial observations.

References

1. Elkinton JR, Singer RB, Barker ES, Clark JK. Effects in man of acute experimental respiratory alkalosis and acidosis on ionic transfers in the total body fluids. J Clin Invest. 1955;34:1671–1690. https://doi.org/10.1172/JCI103221

2. Barker ES, Singer RB, Elkinton JR, Clark JK. The renal response in man to acute experimental respiratory alkalosis and acidosis. J Clin Invest. 1957;36:515–529. https://doi.org/10.1172/JCI103449

3. Polak A, Haynie GD, Hays RM, Schwartz WB. Effects of chronic hypercapnia on electrolyte and acid-base equilibrium. I. Adaptation. J Clin Invest. 1961;40:1223–1237. https://doi.org/10.1172/JCI104353

4. Abdulnour-Nakhoul S, Hering-Smith K, Hamm LL, Nakhoul NL. Effects of chronic hypercapnia on ammonium transport in the mouse kidney. Physiol Rep. 2019;7:e14221. https://doi.org/10.14814/phy2.14221

5. González SB, Menga G, Raimondi GA, Tighiouart H, Adrogué HJ, Madias NE. Secondary response to chronic respiratory acidosis in humans: a prospective study. Kidney Int Rep. 2018;3:1163–1170. https://doi.org/10.1016/j.ekir.2018.06.001

6. Krapf R, Beeler I, Hertner D, Hulter HN. Chronic respiratory alkalosis: the effect of sustained hyperventilation on renal regulation of acid-base equilibrium. N Engl J Med. 1991;324:1394–1401. https://doi.org/10.1056/NEJM199105163242003

7. Schwartz WB, Hays RM, Polak A, Haynie GD. Effects of chronic hypercapnia on electrolyte and acid-base equilibrium. II. Recovery, with special reference to the influence of chloride intake. J Clin Invest. 1961;40:1238–1249. https://doi.org/10.1172/JCI104354

8. Schlichtig R, Grogono AW, Severinghaus JW. Human PaCO₂ and standard base excess compensation for acid-base imbalance. Crit Care Med. 1998;26:1173–1179. https://doi.org/10.1097/00003246-199807000-00015

9. Cadoux-Hudson TA, Rajagopalan B, Ledingham JG, Radda GK. Response of the human brain to a hypercapnic acid load in vivo. Clin Sci. 1990;79:1–3. https://doi.org/10.1042/cs0790001

10. Kety SS, Schmidt CF. The effects of altered arterial tensions of carbon dioxide and oxygen on cerebral blood flow and cerebral oxygen consumption of normal young men. J Clin Invest. 1948;27:484–492. https://doi.org/10.1172/JCI101995

11. Adrogué HJ, Rashad MN, Gorin AB, Yacoub J, Madias NE. Assessing acid-base status in circulatory failure: differences between arterial and central venous blood. N Engl J Med. 1989;320:1312–1316. https://doi.org/10.1056/NEJM198905183202004

12. McKeever TM, Hearson G, Housley G, et al. Using venous blood gas analysis in the assessment of COPD exacerbations: a prospective cohort study. Thorax. 2016;71:210–215. https://doi.org/10.1136/thoraxjnl-2015-207573