Civetta, Taylor, & Kirby's: Critical Care, 4th Edition

Section IV - Essential Physiologic Concerns

Chapter 42 - Blood Gas Analysis and Acid-base Disorders

Steven G. Achinger

Juan Carlos Ayus

Acid-base physiology is among the most complex topics in clinical medicine. Disturbances of this system are common in the critically ill, and important clinical decisions based on measured acid-base parameters occur on a daily, even hourly basis. Therefore, a sound understanding of acid-base physiology is mandatory for the intensivist.

The field is full of complicated concepts and equations that, at times, have only limited applicability to the practicing clinician due to the failure of any current model to faithfully and completely recapitulate the complex buffering process in vivo. The purpose of this chapter is to provide a conceptual introduction to the current approach to acid-base physiology, while de-emphasizing calculations and formulas. The goal is not to know how to derive the commonly used formulas, but rather to understand the meaning of measured and derived acid-base parameters that are used clinically, and how they may—or may not—help in answering three essential questions in the critically ill patient with an acid-base disturbance:

· What acid-base disorder(s) is (are) present?

· How severe is the disturbance?

· What is the etiology underlying the derangements?

Maintenance of the Arterial pH and Acid-Base Balance: Buffering and Acid Excretion

Buffering of acids is the first line of defense against perturbations in systemic pH. Recall that the pH is a logarithmic scale that is a function of the concentration of H₃O⁺ species in solution (H⁺ will be used interchangeably with H₃O⁺ in this chapter). In neutral solution, [H⁺] is × 10^-7 M and [OH^-] is 10^-7 M; this satisfies the dissociation constant for water:

The pH, defined by Sorenson, is the negative log of the concentration of H⁺. Therefore, in neutral solution, the pH is 7. This is a very small concentration of [H⁺], and therefore, addition of small amounts of [H⁺] to water will lead to significant fluctuations in pH. For example, we will consider an experiment performed by Jorgensen and Astrup (1) in which 1.25 mEq/L of HCl is added to hemolyzed human blood. Assuming that the blood contained no buffers (i.e., if the blood was imagined to be a container of water that starts at a neutral pH), the expected pH following such an infusion would be calculated by the following:

1.25 × 10^-2 moles (number of moles of H⁺ added)

+ 1 × 10^-7 moles (number of moles of H⁺ in neutral water)

= 1.25 × 10^-2 moles/ L.

Taking the negative log yields a pH of 1.9; this would be the expected pH if there were not buffers available. However, following the infusion, the pH of the blood changed approximately 0.2 pH points. This means that less than 1/10,000th of the H⁺ added remains unbound in the blood. This illustrates the tremendous buffering capacity of the blood. A buffer can be thought of as a substance that, when present in solution, takes up [H⁺] and therefore resists change in pH when [H⁺] is added. The overall buffering system of the body is complex and includes several components. These are listed in Table 42.1. The most important system is the carbon dioxide–bicarbonate system, which is the principal buffer in the extracellular fluid. This buffering system is also very important clinically since it is the only buffering system where the two components (acid and conjugate base) are readily measurable in the extracellular fluid (ECF). Buffers work by binding the free H⁺ as the conjugate base, which is a weak acid.

Buffers allow the body to resist acute changes in pH; however, buffering capacity will eventually be depleted if acid is continually added. For example, in humans, the net fixed acid production is approximately 70 to 100 mmol/day. It is the excretion of the daily acid load that ultimately allows the body to maintain acid-base balance. The excretion of the daily acid load occurs through two distinct mechanisms: (a) the renal excretion of fixed acid and (b) the respiratory excretion of volatile acid (i.e., carbon dioxide). Through the interconversion of bicarbonate, carbonic acid, and carbon dioxide, fixed and volatile acids can be buffered until they can be excreted through the urine or respiration (Fig. 42.1). In the lung, CO₂ is released, which ultimately leads to more H⁺ reacting with HCO₃^- to generate water and more CO₂. In the kidneys, the entire filtered load of bicarbonate is—in order to avoid losing base—reabsorbed. When the kidney excretes one H⁺ in the urine, one “new” HCO₃^- is generated. These two processes are both important in the excretion of the acid load, and modulation of these processes is also important in compensating for acid-base disturbances.

Urinary Excretion of Fixed Acids

In the reabsorption of bicarbonate, the corresponding H⁺ produced in the process must be excreted in the urine. As most bicarbonate reabsorption occurs in the proximal tubule, the renal secretion of H⁺ is ten times greater in the proximal tubule—approximately 4,000 mmol/day—as compared with the distal tubule—approximately 400 mmol/day. However, in the distal tubule, there is a much higher luminal–intracellular H⁺ gradient than that seen in the proximal tubule—a ratio of approximately 500:1. This high gradient is due to active secretion of H⁺ into the tubule. If the excretion of acid occurred in the absence of buffers, the ability to excrete acid in the urine would be limited. In much the same way that the body can absorb large amounts of acid without much change in pH, the kidney accomplishes a similar task in the excretion of large amounts of fixed acid through the use of buffers in the urine with modestly acidic pH (approximately 5.5 under maximal conditions). The excretion of H⁺in the urine occurs with different conjugate bases, which are grouped as titratable acids—mostly phosphates—and nontitratable acids—ammonium. The excretion of titratable acid has a limit that is, for the most part, dependent on the filtered load of phosphate, as this is the main buffer for nontitratable acids. However, the kidney can generate its own buffer—ammonia; moreover, the renal capacity to generate ammonia and to excrete acid as ammonium under normal conditions is substantial. This capacity may be significantly up-regulated in the face of systemic acidosis. Ammonia is produced in the kidney, traverses the plasma membrane, and is “trapped” in the tubular lumen because the low pH drives the following reaction to the right, as the plasma membrane is much less permeable to the charged species ammonium.

Table 42.1 Blood and extracellular fluid buffers

Acid Conjugate base H2CO3 HCO3– Albumin-H Albumin– H2PO4 HPO4– Hgb-H Hgb–

Therefore, ammonia–ammonium acts as a urinary buffer system, allowing the elimination of one H⁺ for nearly every ammonia produced. The buffering of acid in the urine, especially via ammonium, allows for substantial amounts of acid to be excreted without generating excessively acidic urine.

Figure 42.1. Normal acid-base homeostasis.

Titratable acids make up a relatively small proportion of the acid excreted and do not increase to near the degree that nontitratable acidity (ammonium) increases in the face of systemic acidosis.

The kidney, through active secretion of H⁺ in the distal tubule, is able to achieve an H⁺ concentration gradient of approximately 100:1 between the urine and the intracellular space of the tubular epithelial cells. This corresponds to the maximally acidic urine of approximately pH 5. Without any buffers, it would require 7,000 liters of urine to excrete a daily load of 70 mEq of acid in buffer-free urine of pH 5.0! Therefore, urinary buffers are very important in allowing the body to excrete the daily fixed acid load. Chronic metabolic acidosis stimulates the renal production of ammonia as a physiologic response; this response reaches its maximum production after several days.

Assessing Urinary Acid Excretion

In the presence of systemic acidosis, the kidney will compensate by increasing the excretion of fixed acids, mainly in the nontitratable form (i.e., ammonium). The increase in ammonium—which is a cation; recall that ammonia is predominantly in the form of NH₄⁺ at a pH of 7 or below—excretion results in a perturbation in the electrolytes present in the urine. This manifests as a change in the urine anion gap where ammonium is the unmeasured anion. The urinary anion gap is a useful clinical test that can be used to gauge the amount of ammonium excreted in the urine without directly measuring it (2). It may be used to indirectly estimate the amount of ammonium in the urine and is calculated using the following formula:

The urine anion gap is assessed in patients with metabolic acidosis and is used to determine if the renal response to the systemic acidosis is appropriate. In other words, the urine anion gap answers the question, Are the kidneys excreting the acid load appropriately or are the kidneys part of the acid-base problem? If the kidneys are excreting the acid load appropriately, there must be a nonrenal source of the acidosis—for example, diarrhea. Because ammonium is not a measured cation in this equation, the presence of significant amounts of ammonium causes an abundance of chloride relative to the measured cationic constituents of the urine (sodium and potassium). Therefore, if there is a high level of ammonium in the urine, the urine anion gap will be negative. The relationship between the amount of ammonium in the urine and the urine anion gap is illustrated in Figure 42.2. As a general rule, a negative urinary anion gap suggests that the kidney is excreting ammonium in the urine. This is a continuous variable, and the more negative the value, the greater the renal response. In the face of systemic acidosis, if the renal response is appropriate, there will be a high amount of ammonium in the urine, and the urine anion gap will be highly negative. This is sometimes referred to as a negative net urinary charge; however, this is a bit misleading because the urine is, of course, electroneutral; it is simply because we are not considering the contribution of ammonium that the net urinary charge seems negative. A highly negative urine anion gap is strong evidence that the renal response to metabolic acidosis is normal. Conversely, an anion gap that is near zero or positive suggests that there is little or no ammonium in the urine, which is reflected by a paucity of chloride in the urine relative to the concentration of measured cations. This is evidence that the kidney is not appropriately excreting ammonium in the urine and suggests a renal contribution to the acidosis. A caveat that is often clinically important is that the presence of unmeasured anions in the urine (such as β-hydroxybutyrate) may falsely depress the urinary anion gap, and therefore this test does have some limitations, such as during ketonuria.

Figure 42.2. Effect of urine ammonium concentration on the urine anion gap.

Relationship between Systemic pH and Buffer Concentrations: An Evolving Concept

We have previously noted that, because of the presence of buffers, addition of—or conversely, removal of—[H⁺] to the body does not produce the expected change in pH that would occur in unbuffered solutions. In this section, we will address the question of the relationship between pH and the concentrations of buffers.

Traditional Paradigm

One of the earliest observations in this field, and still very important clinically today, was the observation by Hendersen that the concentration of H⁺ in the blood was dependent upon the concentration of CO₂, H₂CO₃, and HCO₃^-. The Henderson-Hasselbalch equation was later derived by using the Sorenson convention of expressing [H⁺] as pH. This relationship is usually expressed as:

pH = 6.1 + log[HCO₃^-]/[H₂CO₃]

As the [H₂CO₃] in plasma is related to the partial pressure of CO₂, this relationship can be rewritten as:

This relationship became very meaningful clinically as the methodologies to measure the key variables pH, HCO₃^-, and PCO₂ were developed. Now the concentration of the constituents of the principal buffering system could be related to the systemic pH. This allowed, among other things, a framework around which to understand how much alkali must be added in order to affect systemic pH in an acidemic patient. However, it became apparent that the relationship between pH, HCO₃^-, and PCO₂ failed to completely describe the relative contributions of fixed acids and volatile acids (the respiratory component) to acidosis. This is because PCO₂ and HCO₃^- are not truly independent of each other, as changes in one will lead to changes in the other, as will be seen later, according to the relationship given in Eq. 11. Additionally, it was noted that no single value accurately quantifies the degree of fixed acids present during metabolic acidosis or alkalosis. The degree of acidemia could be considered as simply the pH, as the pH is ultimately a composite of the net respiratory and metabolic components of the acid-base disturbance. However, because of the buffering capacity of the body, a quantification of the fixed acid derangement is not explained by this relationship alone.

Historically, several theoretical frameworks have been developed in an attempt to overcome this lack of exactitude in the concepts of quantification and etiology. The first obstacle to accurate quantification is the reality that the HCO₃^- system was not the only quantitatively important buffering system to be considered. The erythrocyte membrane is permeable to H⁺, and therefore H⁺ can diffuse inside the cell and hemoglobin can act as an intracellular buffer. Other buffering systems, such as phosphate and circulating proteins, can act as clinically relevant buffers as well (Table 42.1). By quantitative chemistry, the pH of a system is dependent on the relative concentrations of the acids and conjugate bases of all of the buffering systems present. Clinically, we measure accurately the concentration of the acid (CO₂) and conjugate base (HCO₃^-) of only one buffering system. Therefore, perturbations in the other buffering systems are not accounted for in the framework that only considers carbonate species.

Another major complicating factor is the fact that the human body is not a closed system. CO₂ is both continually being generated in the tissues and continually excreted through the respiratory system. Therefore, changes in CO₂ can, and frequently do, occur very rapidly in humans as the respiratory rate increases or decreases. Additionally, the kidney can modulate the production of HCO₃^- to adjust the HCO₃^- concentration and, albeit at a much slower pace than respiratory effects, change the pH. If the rate of HCO₃^- production exceeds consumption of HCO₃^-, the serum bicarbonate increases; conversely, if it is below the rate of production, HCO₃^- will decrease. What this means is that the concentration of the measured parameters—HCO₃^- and PCO₂—are not just dependent on the inciting insult—the disease process—that caused them to change, but also on the body's response to that change (i.e., compensation).

Base Excess and Standard Base Excess

The change in pH of a system is dependent on both the amount of acid (or base) added and that present on the buffering capacity. As acid is added to a buffered solution, for every H⁺ that is buffered, one molecule of conjugate base of the buffer is consumed. Therefore, assessing changes in concentrations of the conjugate base is more helpful than the degree of change in the pH in quantifying the degree of fixed acids present. The difficulty in describing the buffering system of a patient is the inaccessibility for measurement of a fair proportion of the buffers (Table 42.1), especially intracellular buffers. Several expressions have been proposed to quantify the degree of acid loading based on the change in body buffers. The most commonly used concept in this regard is the base excess. Siggaard-Andersen defined the base excess of blood as the number of mEq of acid (or base) needed to titrate 1 L of blood to a pH of 7.4 at 37° C with a PCO₂ of 40 mm Hg; note that this is an experimentally arrived upon value. The standard base excess is the base excess corrected for changes in hemoglobin, recalling that hemoglobin is an important intracellular buffer. The base excess can be considered as a measurement of the “metabolic” portion of an acid-base disturbance since the concentration of CO₂ is being held constant at a normal level. The base excess has become a widely used parameter to characterize acid-base disturbances. One major drawback of the base excess is that it is a measured parameter of whole blood. However, in vivo, the blood is circulating and comes into contact with other tissues that can serve to provide buffering capacity. In clinical practice, however, the base excess is not measured by titration; rather, it is calculated from a nomogram that assumes normal nonbicarbonate buffers. This simplification, while allowing the widespread application of the base excess, has, in one sense, the drawback of losing the actual measurement of nonbicarbonate buffers that occurs when blood is directly titrated. Despite potential drawbacks, the base excess is very useful in describing the magnitude of a metabolic disturbance on the concentration of buffers and has become a widely used parameter to assess the degree of a metabolic disturbance.

The Anion Gap

The anion gap is calculated by taking the difference between the concentrations of the measured cations and the measured anion; it takes on a value of approximately 8 to 12 mEq/L in healthy individuals (3,4,5). The anion gap is an indirect estimation of the amount of “extra anions” in the circulation. The anion gap normally reflects the serum albumin (negatively charged at physiologic pH) (6), phosphate, and other minor anions (7). The unmeasured anions that may, under pathologic situations, lead to an increased anion gap can be either endogenous substances normally found in lower levels such as lactate or β-hydroxybutyrate or exogenous substances such as salicylates. The anion gap is calculated using the following formula:

Metabolic acidosis is subdivided into anion gap and non–anion gap metabolic acidosis based on the value of the anion gap. In general, metabolic acidosis is caused either by the loss of bicarbonate—as in gastrointestinal losses or impaired renal acid excretion—or by a gain of acid associated with an unmeasured anion. The gain of acid is usually associated with the presence of an unmeasured anion (e.g., lactic acid); an exception might be intake of HCl. The extra base present—again, using the example of lactate—leads to a greater difference between the measured anions and cations, and therefore a greater anion gap. There is a wide range for the normal anion gap (4) and, in our experience, a normal anion gap is approximately 8 to 10 mEq/L—slightly lower than the value referenced above—but this may vary with methodologies in various labs; thus, checking with the local laboratory is imperative (4,8). When interpreting the anion gap, caution must be exercised, as there is significant variation in the anion gap and it can be influenced by many conditions other than metabolic acidosis.

Hypoalbuminemia is the most common condition that affects the normal anion gap since albumin normally contributes to the net negative charge of the blood (9). For every 1 mg/dL fall in the plasma albumin concentration, the anion gap should decrease by approximately 3 mEq/L (3). In plasma cell dyscrasias, such as multiple myeloma, the presence of cationic proteins in the serum is a cause for falsely depressing the anion gap (10), which has been attributed to an increased net positive charge due to the presence of net cationic immunoglobulins (11). Conditions that have been noted to increase the anion gap in the absence of metabolic acidosis are renal failure, volume depletion, metabolic alkalosis (12), and some penicillins. The anion gap can be lowered by hypoalbuminemia, hypercalcemia, and hyponatremia (13). Because of the wide variation in the anion gap and the variety of conditions that can affect it, it is best to directly measure “unmeasured anions” such as lactate whenever feasible.

Case Scenario #1: Use of Henderson-Hasselbalch Equation to Guide Ventilation

A 48-year-old morbidly obese patient is admitted to the hospital with shortness of breath and fever. In the emergency room, he is started on intravenous antibiotics. Over the next 3 hours, he becomes severely short of breath and develops a diminished level of consciousness. He is intubated and placed on mechanical ventilation. His past medical history is significant for diabetes mellitus and hypertension. Social history is significant for one pack per day tobacco abuse for 20 years. Current medications include amlodipine 5 mg PO daily, enalapril 5 mg PO bid, and hydrochlorothiazide 12.5 mg PO bid. Physical exam shows blood pressure of 156/88 mm Hg, pulse 76 beats/minute, and temperature 96°F. The patient is morbidly obese. Cardiovascular exam is normal. Lung exam reveals bilateral breath sounds with diffuse crackles on the right and egophony. The initial ventilator settings are synchronous intermittent mandatory ventilation (SIMV) with a rate of 20, tidal volume of 800 mL, and positive end-expiratory pressure (PEEP) of 5 cm H₂O, with an FiO₂ of 1.0. Thirty minutes after mechanical ventilation is initiated, the following labs are drawn:

Serum
Sodium (mEq/L)	141
Potassium (mEq/L)	4.2
Chloride (mEq/L)	100
Bicarbonate (mEq/L)	34
Blood urea nitrogen (BUN) (mg/dL)	13
Phosphorus (mg/dL)	3.8
Creatinine (mg/dL)	0.8
Albumin (g/dL)	3.8
Glucose (mg/dL)	152
Arterial blood gas
pH	7.65
PO2 (mm Hg)	340
PCO2 (mm Hg)	32

· What acid-base disorder is present in this patient?

This patient has an underlying respiratory acidosis with compensation (note elevated HCO₃^-). When a patient with chronic respiratory acidosis and appropriate renal compensation is placed on mechanical ventilation, he or she is at risk of developing severe alkalemia. This occurs because mechanical ventilation can remove PCO₂ from the blood quickly, hence increasing the pH precipitously. However, it takes time for the kidney to adapt to the change in blood pH. In time, the kidney can adapt by decreasing bicarbonate reabsorption, leading to loss of bicarbonate in the urine if the patient is not chloride depleted, but this does not happen in the acute setting. Following the start of mechanical ventilation in this patient, he has developed an iatrogenic respiratory alkalosis and a dangerously high arterial pH.

· In order to correct the pH to 7.35, what goal CO₂ should be maintained?

The appropriate measure is to decrease the minute ventilation to allow the PCO₂ to rise to a level that would lead to a normal or slightly acidic pH. To determine the PCO₂ that corresponds to a pH of 7.35, use the Henderson-Hasselbalch relationship. In the acute setting, the HCO₃^- will not change since renal adjustments take several days to have full effect, and therefore the best way to change the pH is to adjust PCO₂.

7.35 = 6.1 + log (34/0.03 * PCO₂)

PCO₂ = 64 mm Hg

Therefore, the ventilation rate should be decreased to maintain a PCO₂ of approximately 64 to achieve a pH of 7.35.

Newer Models of Acid-base Quantification: Stewart Approach

The assumptions made in the traditional model are that acids behave as Brønsted/Lowry acids—that is, proton donors—and that the degree of a metabolic disturbance causes a decrease in buffers, which is best approximated by the decrease in serum bicarbonate. Therefore, every mole of H⁺ added results in a reciprocal decrease in the concentration of buffers. This decrease in buffers occurs principally as a decrease in serum bicarbonate, but other unmeasured buffers, such as Hgb-H, are also decreased during acidosis. An approach that is gaining popularity, especially among critical care physicians, is the Stewart model, a deviation from the traditional approach to acid-base quantification, which makes different assumptions about the definition of acids and bases. In essence, in the Stewart model, an acid is defined as any substance that raises the [H⁺] of a solution, not necessarily limited to an H⁺-donating species. Many excellent reviews have been written on this topic (14,15,16,17); discussion in this chapter will be limited to an introduction to the key derived parameters of the Stewart model so that they can be contrasted with the standard approach. The most strikingly different concept of the Stewart model is that the serum bicarbonate is not used as the measure of buffering capacity present. This model uses the strong ion difference as a fundamental measure of the presence of buffers.

The Strong Ion Difference

Strong ions are the ions in the blood that can be considered as completely dissociated in solution (18). The strong ion difference (SID) is analogous to the buffer base of a solution. The SID is calculated as:

The remaining anions in solution are the buffers, which can be thought of as the bicarbonate plus the nonbicarbonate buffers, denoted [A^-]. [A^-] is the sum of negative charge (buffering capacity) of albumin and phosphate.

SID in the Stewart model is considered more reflective of the concentrations of buffers and not the serum bicarbonate. In fact, the Stewart equation describes the pH in terms of three independent variables: The strong ion difference [SID], [A_tot], and PCO₂, where A_tot is the concentration of weak acids. This is in contrast to the Henderson-Hasselbalch equation, which relates pH to [HCO₃^-] and PCO₂(Eq. 4). In the Stewart model, bicarbonate concentration varies dependently on these other more fundamental parameters. The arguments for and against this claim are many, and are beyond the scope of this text. It can be stated, however, that the traditional approach to acid-base disturbances is still practiced most frequently, and the Stewart model has gained widest acceptance in the critical care field. This makes sense in that the Stewart model may have advantages in description of extreme acid-base conditions, especially when the assumption that noncarbonate buffers are constant may not be true, such as in critical illness. A high SID denotes metabolic alkalosis, and a low SID denotes metabolic acidosis. There are modifications of the standard model that attempt to take into account perturbations in noncarbonate species such as the correction of the anion gap for disturbances in serum albumin; this will be illustrated later in examples.

Expected and Apparent Strong Ion Difference

The SID under normal conditions can be thought as the sum of the buffer anions (bicarbonate and nonbicarbonate buffer anions) and should be about 40 mEq/L. As noted above, when the SID deviates from this value, a metabolic acid-base disturbance should be suspected. As noted above, the SID is one of three independent variables that determines [H⁺], and therefore, conversely, the SID can be related to the three fundamental values: pH, PCO₂, and [A_tot]. The expected SID (SIDe) is the SID that would be predicted based on the pH, PCO₂, and A_tot (in this case A_tot is approximated based on the albumin and phosphate). This relationship is given below:

The apparent SID (SIDa) is the strong ion difference considering the concentrations of the strong ions that are normally present in the serum: Na⁺, K⁺, and Cl^-. This definition is given below:

When the SIDa and SID differ, there is a strong anion gap, which is described below.

Strong Ion Gap

The strong ion gap (SIG) should be considered as an evaluation of unmeasured anions, analogous to the traditional anion gap. The strong anion gap is normally zero and is defined as:

SIG = anion gap - [A^-]

where A^- is the composite of nonbicarbonate buffers in the blood. [A^-] = 2.8 (albumin in g/dL) + 0.6 (phosphate in mg/dL)

at a pH of 7.4.

When the strong ion gap exceeds zero, there is an unmeasured anion present. This is analogous to the traditional anion gap, with a correction factor for the presence of hypoalbuminemia (19). The traditional anion gap is rarely corrected for disturbances in phosphate, although, as can be seen in the above formula for A^-, the contribution of deviations in phosphate is much smaller than that of albumin, reflecting that, quantitatively, albumin has much greater buffering capacity than phosphate.

The Stewart model has also been used to classify metabolic acid-base disorders based on the SID. An elevated SID is consistent with metabolic alkalosis, and a low SID is consistent with metabolic acidosis. The metabolic acidoses are further subdivided based on a high SIG (analogous in many ways to a high anion gap) and a low SIG (analogous to a low or normal anion gap). In this regard, the two approaches approximate each other. The use of the SIG may be advantageous over the use of the standard anion gap, given that the anion gap can have a wide range of values and is thus somewhat imprecise. This is especially true in settings where nonbicarbonate buffers deviate from normal—for example, the patient with acidosis, sepsis, and acute renal failure with serum albumin of 1.9, phosphorus of 7.0, and hemoglobin of 7.2 mg/dL. Clearly in this extreme, the assumption that only changes in serum bicarbonate species reflect the metabolic component of the acidosis may not hold true.

Stewart versus Traditional Approach

Despite the differences in these two approaches to acid-base quantification presented, it should be noted that they are quite similar. The advantage of the Stewart approach is that nonbicarbonate buffers are considered in quantifying acid-base disturbances and, as noted before, this is most likely to be pivotal in critical illness. However, accurate quantification of acid-base status is only part of managing acid-base disturbances. Correctly diagnosing acid-base disorders and treating them appropriately is the ultimate goal; in this regard, we do not find considerable advantage of the Stewart approach over more traditional methodologies. It is important that the clinician understand the limitation of any of the acid-base models, such as understanding when perturbations in the anion gap are significant and when they are not. In the cases presented in this chapter, we have used a traditional approach to acid-base analysis, and we continue use this approach in our own practice.

Key Points

1. The buffering of acids allows the body to “absorb” large amounts of acid without significant disturbance in pH. It is through the excretion of the daily acid load, however, that the body is allowed to maintain acid-base balance. A highly negative urinary anion gap suggests that there is significant ammonium in the urine; in response to metabolic acidosis, this would indicate that the renal compensation is intact. A urinary anion gap that is near zero or positive suggests that there is little or no ammonium in the urine and, in the face of systemic acidosis, would indicate renal acid wasting.

2. Bicarbonate is the principal buffer in the body; however, other buffers play an important role in maintaining systemic pH, and disturbances in nonbicarbonate buffers may be more important in critical illness than in other settings.

Volatile Acidity

Up until now, we have dealt exclusively with fixed acids. Disturbances in fixed acids cause a change in available buffers and change in systemic pH. It is critical to note that volatile acidity plays a very important role in determining the systemic pH both in primary respiratory disturbances and in compensation to metabolic disturbances as will be discussed.

Carbon dioxide is soluble in water, and in solution, reacts with water molecules to form carbonic acid, which can then further react as the following:

H₂CO₃ is the acid portion of the bicarbonate buffer; however, its concentration is proportional to the partial pressure of CO₂, and therefore its direct measurement is not necessary. By the equilibrium expressions above, it can be seen that primary changes in PCO₂ will alter the amount of H₂CO₃. A high PCO₂ will increase the concentration of H₂CO₃, and a low PCO₂ will decrease the concentration of H₂CO₃. By quantitative chemistry, as you increase the concentration of the conjugate acid, the pH of a buffered solution will increase, and as you decrease concentration of conjugate acid, the opposite occurs. The PCO₂ in the circulation is the sum of its production and excretion. The production of CO₂ is not frequently altered; however, the excretion of CO₂—occurring only through respiration—is variable based upon the minute ventilation.

Mechanisms of Compensation

The previous sections have detailed how buffering allows the body to absorb significant amounts of H⁺ without large fluctuations in pH. These buffering systems allow pH to remain constant during physiologic changes in endogenous acid production, and they also form the first defense against an acid-base insult (Table 42.1). The buffering systems and respiratory compensation are very rapid, whereas renal compensation may take days to become fully effective (20). The capacity of the body's buffering system is substantial, and therefore significant amounts of acid can be absorbed before a relatively small change in systemic pH occurs. Buffers act immediately and are thus the first line of defense.

In response to systemic changes in pH, compensatory mechanisms act to counteract the primary disturbance. Figure 42.3 gives an overview of the compensatory responses to the primary acid-base disturbances. In response to metabolic acidosis, there is increased ventilation to decrease PCO₂; the kidney responds by increasing the excretion of H⁺, thereby generating more HCO₃^- (Fig. 42.3A). The opposite response occurs during metabolic alkalosis, except that the kidneys are usually not able to respond to the increase in HCO₃^- appropriately, as failure of the kidney to respond to elevated HCO₃^- is necessary for the development of metabolic alkalosis (this is described later) (Fig. 42.3A). The response to elevated PCO₂ is to increase the renal excretion of H⁺, which leads to increased HCO₃^- production (Fig. 42.3B). In response to metabolic alkalosis, the excretion of H⁺ decreases and less bicarbonate is produced, thereby decreasing serum bicarbonate (Fig. 42.3B). Note that each of the four primary disturbances leads to a perturbation of one of the carbonate species. Consequently, the main compensatory response is to alter the concentration of the conjugate species so that the ratio between the two can be maintained. The response to respiratory disorders has acute and chronic components. The acute response is related to immediate buffering, and the chronic phase of the response occurs as renal compensation takes place; only the chronic phase is depicted in Figure 42.3B. An important point to remember in determining the acid-base disturbances present in a patient is that a compensatory response will never normalize the serum pH or lead to a recovery of the pH past neutrality. If this has occurred, there must be another acid-base disturbance present.

Figure 42.3. A: Primary metabolic acid-base disturbances and compensatory mechanisms. B: Primary respiratory acid-base disturbances and compensatory mechanisms.

Figure 42.4. Approach to the critically ill patient with an acid-base disorder.

Determining Which Acid-Base Disturbances Are Present

A systematic approach is important in correctly diagnosing acid-base disorders in critically ill patients. Because it is common to have mixed acid-base disorders with two or even three disorders present, it is important to evaluate the available information thoroughly and avoid quick judgments based on an incomplete picture.

The approach to acid-base disorders is summarized in Figures 42.4 and 42.5. A key is to identify the primary disturbance. This is best accomplished by analyzing acid-base data in conjunction with a good history. The physical examination is rarely helpful in determining the etiology of an acid-base disorder. It is also important to note that the algorithm in Figure 42.5 is useful in the case of single acid-base disorders. Mixed disturbances are discussed later. Once the primary disturbance is identified, the next step is to assess the adequacy of compensation.

Table 42.2 gives the expected values of PCO₂ and HCO₃^- following compensations for primary disturbances. In the setting of respiratory disorders, acute compensation occurs over hours, while the chronic compensation occurs over days. If there is only one disturbance, and the compensation is adequate, there is nothing more to do. On the other hand, if the compensation is inappropriate, then a second disorder is present. Recall that a compensation will never normalize the pH or “compensate” past the point of neutrality (e.g., a primary metabolic acidosis as a single disorder cannot lead to a neutral or alkaline pH). In these cases, a mixed acid-base disorder is present. An additional clue that a mixed acid-base disorder is present is an elevated anion gap when a metabolic acidosis is not suspected. For this reason, it is good practice to calculate the anion gap in all critically ill patients.

Table 42.2 Compensations for acid-base disorders

Metabolic acidosis · For every 1 mmol/L decrease in HCO3– → 1 mm Hg decrease in PCO2 · Expected PCO2 = 1.5 (HCO3–) + 8 ± 2 · PCO2 should approach last two digits of pH Metabolic alkalosis · For every 1 mmol/L increase in HCO3– → 0.7 mm Hg increase in PCO2 Respiratory acidosis · Acute: For 10 mm Hg increase in PCO2 → 1 mmol/L increase in HCO3– · Chronic: For 10 mm Hg increase in PCO2 → 4 mmol/L increase in HCO3– Respiratory alkalosis · Acute: For every 10 mm Hg decrease in PCO2 → 2 mmol/L decrease in HCO3– · Chronic: For every 10 mm Hg decrease in PCO2 → 4 mmol/L decrease in HCO3–

Figure 42.5. General approach to acid-base disorders. (From Ayus JC. MKSAP 14, 2006, Nephrology Section. Medical Knowledge Self Assessment Program. American College of Physicians; 2006. Copyrighted by the American College of Physicians.)

Causes of Acid-Base Disorders

In determining the etiology, the clinician usually must rely on the history, clinical presentation, and, most importantly, laboratory data in order to determine the inciting disease state (Fig. 42.6).

Metabolic Acidosis

Causes of Anion Gap Acidosis

The common etiologies of elevated anion gap metabolic acidosis are listed in Table 42.3.

Diabetic Ketoacidosis

Insulin is secreted and mediates the metabolism of carbohydrates and the storage of fat during times of normal enteral intake. Under fasting conditions, insulin secretion decreases. Diabetic ketoacidosis occurs when a deficit of insulin activity leads to altered cellular metabolism and glucose utilization is impaired. The deficiency of insulin causes the liberation of fatty acids and pathophysiologic keto acid production. The degree of increase in the anion gap is related to the amount of retained ketones, and therefore diabetic ketoacidosis can be present with varying degrees of hyperchloremia (21). In addition to abnormal ketoacid production, diabetic ketoacidosis is typically also associated with hyperglycemia, a decrease in circulatory volume, and, oftentimes, free water deficits as well, in addition to hypokalemia and hypophosphatemia. Even though total body potassium stores are decreased, the serum potassium concentration is frequently elevated on presentation due to the effects of insulin deficiency, hyperglycemia, and acidosis on potassium distribution. The treatment of diabetic ketoacidosis includes re-expansion of the extracellular fluid volume, administration of insulin to halt acid production, and correction of potassium and phosphorus deficits, with close monitoring of plasma electrolytes.

Figure 42.6. Diagnostic approach to metabolic acidosis. (From Ayus JC. MKSAP 14, 2006, Nephrology Section. Medical Knowledge Self Assessment Program. American College of Physicians; 2006. Copyrighted by the American College of Physicians.)

Table 42.3 Common causes of elevated anion gap metabolic acidosis

Lactic acidosis Diabetic ketoacidosis Renal failure Ingestions · Methanol · Ethylene glycol · Paraldehydes · Salicylates

Lactic Acidosis

There are two types of lactic acidosis: Type A lactic acidosis is due to tissue hypoxia and the formation of excess lactic acid, and constitutes the majority of cases of lactic acidosis. This is frequently seen in sepsis, profound anemia, shock, hypotension, and bowel and limb ischemia. Lactate is formed in tissues under hypoxic conditions as it is a by-product of anaerobic cellular metabolism. Type A lactic acidosis is often associated with poor outcomes if the cause is not quickly reversed, usually due to the severity of the underlying condition, such as septic shock or bowel infarction.

Type B lactic acidosis occurs when there is insufficient liver metabolism of lactate. The normal metabolism of lactate leads to the generation of bicarbonate and therefore when this pathway is less operative, there is decreased bicarbonate and systemic acidosis. This condition can be seen in severe liver disease and/or other conditions that interfere with liver metabolism. Several commonly used medications have been associated with lactic acidosis including propofol, metformin, the nonnucleotide reverse transcriptase inhibitors, stavudine, didanosine, and zidovudine. Carbon monoxide poisoning can present with nonspecific symptoms and lead to lactic acidosis by inhibiting oxygen utilization in the tissues.

Toxic Ingestions Associated with Elevated Anion Gap Acidosis

Ingestions are important causes of acidosis in the critical care setting (22). Common ingestions that often lead to elevated anion gap metabolic acidosis are listed in Table 42.3. The presence of ingested alcohols or other solvents can be inferred by measurement of the osmolal gap when an ingestion is suspected.

The osmolal gap can be calculated using the following formula:

(Measured serum osmolality) - (Calculated serum osmolality)

where the calculated serum osmolality is obtained as follows:

The osmolal gap is normally approximately 10 mOsm/kg H₂O. The normal osmolal gap is a reflection of substances normally present in the serum that exert oncotic forces. These are plasma proteins and ions found in smaller quantities, such as calcium and magnesium. An elevated osmolal gap is an indication that an unmeasured osmole is present in the serum; in the intensive care setting, this is commonly due to ethanol. The osmolal gap can also be used to quantify the level of ethylene glycol or methanol, although direct measurement of these toxins is indicated if their presence is suspected; however, therapy should not be delayed while waiting for confirmation.

Case Scenario #2: Salicylate Toxicity

A 68-year-old man presents to the emergency room following an intentional toxic ingestion. He was brought in by his son, who found him confused when he stopped by his (the patient's) house. One month prior to admission, he suffered the loss of his wife and has felt hopeless since that time. Upon presentation, he is lethargic and weak. His past medical history is significant for hypertension and gout. Past surgical history is significant for an appendectomy 20 years ago and coronary artery bypass graft (CABG) 2 years ago. He smokes one and a half packs of cigarettes a day and he denies the use of alcohol or illicit drugs. Physical exam is significant for blood pressure of 156/80 mm Hg, pulse of 79 beats/minute, respirations of 32 breaths/minute, and temperature of 98°F. He is lethargic, in moderate respiratory distress, and oriented only to place and person. Cardiovascular exam is normal, and there is no wheezing on chest exam. Laboratory data are given below:

Serum
Sodium (mEq/L)	141
Potassium (mEq/L)	3.9
Chloride (mEq/L)	105
Bicarbonate (mEq/L)	9
BUN (mg/dL)	21
Creatinine (mg/dL)	1.2
Glucose (mg/dL)	128
Arterial blood gas
pH	7.40
PO2	67
PCO2	15

Salicylate toxicity is typically associated with an anion gap metabolic acidosis and respiratory alkalosis; this is the acid-base disorder present in this patient. In diagnosing the acid-base disturbance in this patient, the first step is to look at the serum pH. The pH is normal with low serum bicarbonate. This is the first clue that a complex disorder is present since a compensatory response to metabolic acidosis would not normalize the serum pH. We can determine that there is a metabolic acidosis present because of the elevated anion gap (anion gap = 27). Given the presence of a metabolic acidosis, we can then predict what the PCO₂ should be. Using the Winter formula, the expected PCO₂ is approximately 21 mm Hg (Table 42.2), and thus a respiratory alkalosis is present. This pattern is strongly suggestive of salicylate toxicity (23).

Case Scenario #3: Lactic Acidosis in the Setting of Abnormal Levels of Nonbicarbonate Buffers

A 44-year-old female with cirrhosis secondary to autoimmune hepatitis is admitted to the hospital for fever and abdominal pain. The patient is listed for an orthotopic liver transplantation and has been clinically stable for the past month. She noted abdominal pain and fever that have gotten progressively worse over the last 2 days. Her past medical history is otherwise nonsignificant. Current medications include spironolactone 100 mg PO bid, furosemide 80 mg PO bid, and lactulose 30 mL PO bid. Previous surgeries include the placement of a transjugular intrahepatic portosystemic shunt (TIPS) and a cholecystectomy. Physical exam is significant for blood pressure of 74/55 mm Hg, pulse of 72 beats/minute, temperature 100.8°F, and respiratory rate of 24 breaths/minute. She appears cachectic. Cardiovascular and chest exams are normal. Her abdomen is distended and there is diffuse tenderness. She has 1+ pitting edema in the lower extremities. Spontaneous bacterial peritonitis is suspected, and the patient is admitted to the hospital. Admission labs are given below:

Serum
Sodium (mEq/L)	128
Potassium (mEq/L)	5.1
Chloride (mEq/L)	106
Bicarbonate (mEq/L)	11
BUN (mg/dL)	20
Creatinine (mg/dL)	1.3
Phosphorus (mg/dL)	2.1
Albumin (g/dL)	1.4
Glucose (mg/dL)	84
Arterial blood gas
pH	7.23
PO2 (mm Hg)	78
PCO2 (mm Hg)	28

· What is the best characterization of the acid-base disturbance in this patient?

The acid-base disorder is an anion gap metabolic acidosis, likely lactic acidosis. While the anion gap appears normal, the degree of hypoalbuminemia needs to be considered because the negative charge on albumin is a significant component of the “normal” unmeasured anions. For every 1 g/dL decrease in the serum albumin, the expected anion gap decreases by 2.5 mEq/L. Thus, in this patient, to consider her anion gap at a normal level, it should not exceed approximately 5 to 6 mEq/L. Again, caution should be used in interpreting the anion gap in these settings; if any suspicion exists for lactic acidosis, the serum lactate should be measured directly.

Table 42.4 Etiologies of non–anion gap metabolic acidosis

Extrarenal source Renal source Gastrointestinal disorders · Diarrhea · Pancreatic and biliary fistulas · Laxatives, cholestyramine · Ureterointestinal diversions (ileal conduit) · Renal tubular acidosis · Renal failure · Hypoaldosteronism

Non–Anion Gap Acidosis

In a non–anion gap/hyperchloremic acidosis, there is a primary decrease in the serum bicarbonate and an associated increase in the serum chloride. The serum bicarbonate decreases because of renal or extrarenal, gastrointestinal losses (Table 42.4). One of the key etiologic questions to be answered in the approach to a patient with a non–anion gap acidosis is whether or not the kidney is appropriately responding to the acidosis by excreting the acid load or if the cause of the acidosis is improper acid excretion by the kidney. This allows one to differentiate renal from nonrenal causes of the acidosis. The urine anion gap, which is calculated using Eq. 3, is a convenient methodology to assess urinary acid excretion. If the urine anion gap is positive or close to zero, this suggests that (a) there is very little ammonium in the urine (Fig. 42.2) and (b) the kidney is not appropriately excreting acids. If the urine anion gap is highly negative, this suggests that (a) there is a large amount of ammonium in the urine and (b) the kidney is excreting acids appropriately (Fig. 42.2).

Causes of Non–Anion Gap Metabolic Acidosis

Diarrhea

Severe diarrhea leads to non–anion gap metabolic acidosis through loss of bicarbonate, and is typically associated with volume depletion and hypokalemia. In very severe cases, circulatory collapse can occur, and an anion gap (lactic) acidosis may supervene upon the non–anion gap acidosis. Patients who chronically abuse laxatives may develop metabolic acidosis and hypokalemia. However, frequently, these patients also abuse diuretics and therefore can have an associated metabolic alkalosis. In order to determine if the renal response to the acidosis is normal, the urine anion gap should be measured.

Ureterointestinal Diversions

In a ureterointestinal diversion, urine in the intestine leads to reabsorption of chloride and water. Consequently, the absorption of chloride can induce secretion of bicarbonate into the intestine. Additionally, urease-positive bacteria in the intestine metabolizes the urea in the urine to form ammonium, which, when absorbed, liberates excess acid after it is metabolized in the liver. Also, chronic pyelonephritis is common in the diverted kidney, and a superimposed distal renal tubular acidosis (RTA) may occur.

Renal Tubular Acidosis

Renal tubular acidosis is a heterogeneous mix of disorders that is characterized by defects in urinary acid excretion in the setting of intact renal function. Proximal (type 2) RTA is caused by a decrease in proximal bicarbonate reabsorption, whereas in distal (type 1) RTA, the primary defect is impairment of distal acidification (24,25). The net result is that the urine pH is not maximally acidified. The lack of acidification—in other words, secretion of H⁺—leads to less ammonium trapping in the urine and therefore to an anion gap that is either positive or near zero. In the intensive care unit, renal tubular acidosis often presents with profound acidosis and hypokalemia. It is important to treat the hypokalemia first with potassium chloride before correcting the acidosis, as administration of bicarbonate in the setting of severe potassium depletion can lead to fatal hypokalemia as potassium is taken up by cells when H⁺ exits the cells. Type 4 RTA is a clinical syndrome of hyperkalemia and hyperchloremic metabolic acidosis (26) caused by a lack of aldosterone effect on the kidney and is seen most commonly in the following settings: diabetes, advanced age, acquired immunodeficiency syndrome (AIDS), interstitial nephritis, obstructive uropathy, postrenal transplant status, use of angiotensin-converting enzyme inhibitors and heparin (both of which impair aldosterone production), and use of cyclosporine.

RTA should be suspected if the renal response to systemic acidemia is impaired as evidenced by a positive urine anion gap. The next step is to determine the type. The most practical starting point is to differentiate a proximal from a distal RTA. To understand this, some physiology must be discussed. Recall that bicarbonate is reabsorbed predominantly in the proximal tubule. Proximal RTA develops because of impaired reabsorption of bicarbonate. The lack of bicarbonate reabsorption in patients with normal serum bicarbonate leads to wasting of bicarbonate in the urine until a steady state is reached in which the serum bicarbonate drops to a level at which the reabsorptive capacity of the proximal tubule is no longer overwhelmed. At this point, there is no longer any bicarbonate in the urine. For this reason, in a patient with proximal RTA, the serum bicarbonate will be low; however, the urine pH will be low—this is because the distal acidification mechanisms are functional. If such a patient is given an alkali load, serum bicarbonate is temporarily increased, and bicarbonate “spills” into the urine because the filtered load of bicarbonate exceeds the reabsorptive threshold, which leads to an increase in the urine pH. Once the alkali load is stopped, serum bicarbonate drops, bicarbonate no longer appears in the urine, and the urine pH can now drop to its maximally acidic level of approximately 5.5. This is the basis for the provocative testing to demonstrate a proximal RTA, and also explains why these patients often have to take a tremendous amount of alkali in order to achieve normal serum pH.

Renal Failure

The kidneys have the capacity to excrete acids to such a degree that acid-base balance is maintained until kidney function deteriorates to below a glomerular filtration rate of approximately 20 mL/minute. The resulting acidosis is of a mixed type and it is generally, but not universally, associated with an elevated anion gap. Chronic metabolic acidosis should be treated to prevent bone demineralization, which may occur with time. The goal of treatment is to maintain normal acid-base status.

Pancreatic or Biliary Fistula

These disorders can lead to the loss of bicarbonate-rich solutions through the gastrointestinal tract and result in systemic acidosis. If correction of the underlying fistula is not possible, treatment with alkali salts can be helpful.

Hypoaldosteronism

Similar to type 4 RTA acidosis, hypoaldosteronism can lead to an impairment in renal acid excretion. Aldosterone activity in the kidney leads to hypokalemia and metabolic alkalosis; conversely, lack of this activity decreases aldosterone secretion and leads to hyperkalemia and metabolic acidosis.

Case Scenario #4: Non–Anion Gap Metabolic Acidosis: Assessing Urinary Acid Excretion

A 66-year-old man is seen in the emergency room. He has had 8 days of severe diarrhea, abdominal pain, and decreased food intake, but adequate intake of liquids. He believes that he became sick after babysitting his grandson who had similar symptoms. His medical history is significant for diabetes and hypertension. Surgical history only consists of coronary artery bypass grafting 3 years ago. His medications include enalapril 20 mg PO bid, aspirin 81 mg PO daily, atenolol 50 mg PO daily, hydrochlorothiazide 25 mg PO daily, and metformin 1 g PO bid. He has a family history of diabetes and premature coronary artery disease. He does not smoke or use drugs, and drinks alcohol occasionally. Physical exam is significant for blood pressure of 105/70 mm Hg and a pulse of 72 beats/minute; blood pressure drops to 90/50 mm Hg when the patient stands. Temperature is 98.8°F, and respiratory rate is 32 breaths/minute. There is a small amount of occult blood in the stool. Labs are given below:

Serum
Sodium (mEq/L)	136
Potassium (mEq/L)	3.9
Chloride (mEq/L)	114
Bicarbonate (mEq/L)	13
BUN (mg/dL)	21
Creatinine (mg/dL)	1.2
Albumin (g/dL)	4.0
Glucose (mg/dL)	128
Urine
pH	6
Sodium (mEq/L)	32
Potassium (mEq/L)	21
Chloride (mEq/L)	80
Arterial blood gas
pH	7.27
PO2	90
PCO2	30

· Which acid-base disorder is present and what is the likely etiology?

This patient has a non–anion gap metabolic acidosis from a nonrenal origin. The low pH and decreased serum bicarbonate indicate the presence of metabolic acidosis. Respiratory compensation is adequate, and therefore there is no complex acid-base disorder present. The serum anion gap is not elevated. The urine electrolytes and the calculation of the urine anion gap are useful to distinguish between a renal source and a gastrointestinal (GI) source of the acidosis. If gastrointestinal losses are the cause of the acidosis and the renal response to the acidosis is normal, a significant amount of ammonium will be present in the urine. The presence (or absence) of ammonium can be inferred by calculating the urine anion gap. The formula for the urine anion gap is as follows: [K⁺] + [Na⁺] – [Cl^-]. If there is an unmeasured anion present, then [Cl^-] exceeds [K⁺] + [Na⁺] and the urine anion gap is significantly negative. When there is little or no unmeasured anion present, the urine anion gap will take on a positive value. In this case, the urine anion gap = 32 mEq/L + 21 mEq/L – 80 mEq/L = -27 mEq/L, and therefore there is a significant amount of ammonium (NH₄⁺) in the urine, which implies a normal renal response to the systemic acidosis—thereby designating an extrarenal cause of the acidosis.

Treatment of Metabolic Acidosis

Treatment of Anion Gap Metabolic Acidosis

The treatment of an anion gap metabolic acidosis is focused on reversing the pathogenesis of the endogenous acid production and eliminating excess acid. By far the most important aspect of treatment is to identify the source of the acidosis if it is not already apparent, such as in diabetic ketoacidosis or septic shock. Treatment of an anion gap acidosis with bicarbonate replacement therapy remains controversial, especially in lactic acidosis. It has been argued that bicarbonate may be used as a bridge until homeostatic mechanisms reverse the condition through the metabolism of endogenous bases, such as lactate and ketone bodies, and therefore bicarbonate regeneration. This approach of using alkali therapy assumes that there is a detriment to a low pH (or serum bicarbonate) above and beyond the harm caused by the underlying condition. However, evidence from animal models argues that bicarbonate therapy may have deleterious effects on pH, serum lactate levels, and cardiac function (27,28,29). Bicarbonate leads to the generation of CO₂ during buffering and, as CO₂ readily diffuses across cell membranes, intracellular acidosis has been shown to worsen during bicarbonate therapy. Worsening of cardiac function, which has been associated with intracellular acidosis, is the proposed mechanism for worsening of lactic acidosis following the administration of bicarbonate during lactic acidosis (30). Hemodialysis rapidly corrects acidosis and is typically necessary to treat acidosis in the setting of renal failure.

Treatment of Non–Anion Gap Metabolic Acidosis

Bicarbonate therapy is generally indicated in non–anion gap acidosis since the primary disturbance is a decrease in bicarbonate. This is contrasted to anion gap acidosis where correction of the underlying cause is the primary concern. Oral bicarbonate or oral citrate solutions are agents for chronic therapy for non–anion gap acidosis. For acute presentations, especially in patients who may not be able to tolerate prolonged hyperventilation, intravenous bicarbonate therapy may be used.

Medications and Metabolic Acidosis

Medications are an increasingly important cause of severe acidosis and can be life threatening in many cases. Lactic acidosis has been reported with all nonnucleoside reverse transcriptase inhibitors used to treat human immunodeficiency virus (HIV); this effect is related to the drug's inhibition of mitochondrial function, with resultant anaerobic metabolism (31,32,33,34). The newer-generation anticonvulsant topiramate has also been associated with lactic acidosis (34). Metformin is also well known to lead to lactic acidosis, which can be treated with hemodialysis (35). The propofol infusion syndrome is a dangerous complication sometimes seen with the use of this drug; it is associated with head injury, use of propofol for more than 48 hours, use in children, and concomitant use of catecholamines and steroids (36,37,38).

Case Scenario #5: Propofol Infusion Syndrome

A 25-year-old male is in the intensive care unit following a craniotomy for a traumatic head injury. He had suffered a depressed skull fracture to the left frontal bone from blunt trauma during an altercation. He has no known medical problems and takes no medications. Family members state that he occasionally uses intravenous

P.643

cocaine and smokes cigarettes. Intraoperatively, he is given intravenous cefazolin and phenytoin. It is now postoperative day 2, and he is currently receiving propofol infusion at 8 µg/kg/minute. Blood pressure is 155/90 mm Hg, pulse 80 beats/minute, temperature 97.4°F, and respiratory rate 12 breaths/minute. He is currently mechanically ventilated on SIMV mode with bilateral breath sounds. He has a normal cardiac exam, and there is no peripheral edema. Laboratory data are as follows:

	Day 3	Day 6
Sodium (mEq/L)	136	137
Potassium (mEq/L)	3.9	4
Chloride (mEq/L)	104	103
Bicarbonate (mEq/L)	20	12
BUN (mg/dL)	18	19
Creatinine (mg/dL)	1.1	1.0
Albumin (g/dL)	4.0	4.0
Glucose (mg/dL)	128	112
Lactate (mmol/L)		7
pH	7.37	7.21
PCO2	38	32

· What is the likely cause of the acidosis in this patient?

Propofol infusion syndrome is an important entity in the intensive care unit (37,39,40). The patients who appear to be at the greatest risk for the condition are those receiving prolonged infusions after suffering brain injury. Treatment for the condition appears to be discontinuation of propofol; hemofiltration has been used successfully (41,42). It is important to note that many fatalities have occurred when the condition is not recognized promptly, and thus early recognition is critical.

Key Points

1. Bicarbonate therapy is indicated for non–anion gap acidosis.

2. The primary concern in anion gap acidosis is correction of the underlying cause.

Metabolic Alkalosis

Metabolic alkalosis occurs when there is an excess of buffers present, raising the systemic pH. In metabolic alkalosis, there is a primary elevation in the serum bicarbonate. This condition is common in the intensive care setting and can have severe complications. As the primary problem is an increase in bicarbonate, metabolic alkalosis can be readily corrected by renal bicarbonate excretion. Under normal circumstances the potential for bicarbonate excretion is tremendous, and thus, alterations in the renal handling of bicarbonate must occur to maintain the alkalosis. Without an impairment of the renal capacity to excrete bicarbonate, the kidneys would simply excrete the bicarbonate load. The most common reason for impairment of renal excretion of bicarbonate is chloride deficiency and renal failure. In general, metabolic alkaloses are generated by either bicarbonate intake in excess of loss or by the primary loss of H⁺ (Table 42.5).

Table 42.5 Causes of alkalosis generation

Excessive alkaline load Loss of hydrogen ions Bicarbonate infusion, hemodialysis Gastrointestinal losses: Vomiting, nasogastric suction CaCO3 supplements Renal losses: Diuretics, excessive Oral citrate solutions mineralocorticoid Parenteral nutrition (acetate, glutamate)

Chloride-sensitive Metabolic Alkalosis

Nasogastric suction, vomiting, and diuretics are very frequent causes of metabolic alkalosis. Hypokalemia develops in the setting of vomiting or nasogastric suction not due to gastrointestinal losses, as the stomach contents are not rich in potassium; rather, the losses of potassium are renal losses due to potassium bicarbonate excretion and secondary hyperaldosteronism. In these settings, renal losses of sodium and potassium are obligatory in order to excrete bicarbonate. In this situation, the urinary chloride (not the urinary sodium) better reflects the effective blood volume of the patient. Similar to the loss of gastric secretions, diuretic-induced extracellular fluid volume depletion stimulates aldosterone secretion. The action of aldosterone stimulates sodium reabsorption in the distal tubule, which is coupled with secretion of potassium and H⁺. Therefore, a urine that is paradoxically acidic is generated. Other causes of metabolic alkalosis that are sensitive to the administration of chloride include those occurring after hypercapnic and after diuretic use.

As noted above, in order to maintain the alkalosis, renal bicarbonate excretion must be impaired in some way. In the setting of chloride depletion, the kidney is unable to excrete the excess bicarbonate, and therefore the alkalosis is maintained (43,44) (Table 42.6). Among patients with normal renal function and normal chloride status, attempting to raise the serum bicarbonate concentration 2 to 3 mEq/L above the normal value is virtually impossible because the kidneys can easily excrete the excess bicarbonate.

Table 42.6 Classification of metabolic alkalosis by chloride handling

Chloride sensitive (urine Cl– less than 20 mEq/L) Chloride resistant (urine Cl– greater than 40 mEq/L) Gastrointestinal acid losses · Nasogastric suction, vomiting · Congenital Cl– losses in stool? · Rectal adenoma Hypertensive · Renovascular hypertension · Hyperaldosteronism · Liddle syndrome · Glycyrrhizic acid Renal acid losses · Penicillins, citrate · Postdiuretic · Posthypercapnic Normotensive · Diuretics · Bartter and Gitelman syndromes · Administration of alkali

Chloride-insensitive Metabolic Alkalosis

The chloride-insensitive metabolic alkalosis commonly encountered in the critical care setting is that occurring after the use of loop diuretics. The loss of large amounts of bicarbonate-free fluid in a patient with expanded ECF space—such as during therapy with a loop diuretic—is thought to lead to a reduction in the ECF space, with relative conservation of bicarbonate concentration. This has been termed contraction alkalosis. Other causes of chloride-insensitive metabolic alkalosis are hyperaldosteronism—both primary and secondary—such as might be seen with renovascular disease (Table 42.6). Rare causes of chloride-insensitive metabolic alkalosis are Bartter and Gitelman syndromes.

Renal and Extrarenal Compensation

Immediately following the generation of metabolic alkalosis, buffering systems begin to decrease the effects of the alkaline load. Respiratory compensation for a metabolic alkalosis involves respiratory suppression and an increase in the PCO₂ (Fig. 42.3A, Table 42.2). Respiratory compensation for severe metabolic alkalosis has practical limits, as respirations can be suppressed only to a certain degree. Without the effect of mitigating factors such as volume depletion, the kidney will respond to metabolic alkalosis through increasing the renal excretion of bicarbonate. Severe chloride depletion can theoretically inhibit this exchange and therefore inhibit bicarbonate secretion. Finally, hyperaldosteronism secondary to diuretic use stimulates the tubular secretion of potassium and H⁺. The net effect is an acidic urine that also helps to maintain the alkalosis. In patients with low urinary chloride, chloride replacement is indicated to allow bicarbonate excretion.

Treatment of Metabolic Alkalosis

The metabolic alkalosis seen in the intensive care unit often develops as a complication rather than a presenting disorder. H₂ blockers and proton pump inhibitors can be used as a measure to decrease losses of H⁺ in patients with prolonged gastric aspiration or chronic vomiting, which may help prevent the development of metabolic alkalosis. In patients with chloride-sensitive metabolic alkalosis, treatment usually consists of replacement of the chloride deficit—usually with normal saline since volume depletion is also often present. Potassium chloride is almost always indicated when hypokalemia is also present, although potassium concentrations may increase as the alkalosis is corrected. In severe, symptomatic metabolic alkalosis—a pH greater than 7.6—hemodialysis may be indicated and can be used to correct alkalemia, especially when associated with renal failure (45). The use of acidic solutions is rarely indicated (Table 42.7).

Case Scenario #6: Diabetic Ketoacidosis with Concomitant Metabolic Alkalosis

A 21-year-old male presents to the emergency room with severely diminished mental status. He states that he has felt nauseated for the last few days and has been unable to eat well. This morning, he vomited several times and was brought to the emergency room by his girlfriend. His past medical history is negative for any chronic medical problems. He had a tonsillectomy as a child but no other surgeries. Physical exam is significant for blood pressure of 122/57 mm Hg, pulse of 105 beats/minute, respiratory rate of 28 breaths/minute, and temperature of 99.3°F. He is thin and in moderate distress. Chest exam is normal. His abdomen is soft and nontender. Stool is negative for occult blood. In the emergency room, the patient begins to vomit large amounts, and he aspirates a significant amount of stomach contents and develops respiratory failure. He is intubated and started on mechanical ventilation. After 1 hour of mechanical ventilation, the following laboratory values are received:

Serum
Sodium (mEq/L)	138
Potassium (mEq/L)	3.7
Chloride (mEq/L)	91
Bicarbonate (mEq/L)	16
BUN (mg/dL)	11
Phosphorus (mg/dL)	2.2
Albumin (g/dL)	3.6
Creatinine (mg/dL)	1.7
Glucose (mg/dL)	980
Arterial blood gas
pH	7.41
pO2	67
PCO2	27

· What is the acid-base disturbance present in this patient?

This patient has a mixed acid-base disorder, metabolic acidosis/metabolic alkalosis. The patient presents with diabetic ketoacidosis. The anion gap is 31, which signifies a large degree of ketoacid production. Because of the nausea and vomiting, he also has developed a metabolic alkalosis, and thus the bicarbonate level is higher than one would expect with this degree of acid production. This can be formalized by calculating the delta–delta anion gap. Another method of conceptualizing what is occurring is to take the difference of the anion gap and a normal anion gap. To illustrate how this works, we define the normal anion gap as 12 mEq/L. In this case, the difference between the patient's anion gap and the normal anion gap is: 31 – 12 = 19 mEq/L. This difference is often referred to as the delta–delta anion gap. If this number is added to the patient's bicarbonate, the result is 35. This significantly exceeds the normal bicarbonate of 24, which indicates that a metabolic alkalosis is present. What this tells us is that if all of the unmeasured anions—which are potentially bicarbonate—are converted back to bicarbonate, the patient would be considered to have a metabolic alkalosis.

Table 42.7 Treatment of metabolic alkalosis

Chloride sensitive · IV normal saline volume expansion · Discontinue diuretics if possible · H2 blockers or proton pump inhibitors in cases of nasogastric suction and vomiting Chloride resistant · Remove offending agent · Replace potassium if deficient Extreme alkalosis · Hemodialysis · NH4Cl or HCl can also be used

Key Points

1. Metabolic alkalosis is often accompanied by a decrease in chloride such that the decrease offsets the incremental increase in bicarbonate.

2. Metabolic alkalosis is caused by excessive bicarbonate intake or loss of H⁺.

3. Vomiting, nasogastric suction, and diuretics are the most frequent causes of metabolic alkalosis in the intensive care unit setting.

4. In patients with metabolic alkalosis and low urinary chloride, normal saline is indicated to expand the extracellular space.

Respiratory Acid-base Disorders

Under normal conditions, through endogenous metabolism, approximately 15,000 mmol/day of CO₂ is produced. Carbon dioxide enters the plasma and forms carbonic acid, which subsequently dissociates to bicarbonate and H⁺. The majority of this CO₂ generated is transported to the lungs in the form of bicarbonate. The H⁺ produced in the process is exchanged across the erythrocyte cell membrane and is buffered intracellularly. In the alveoli, this process is reversed and the bicarbonate combines with H⁺, liberating CO₂, which is then excreted through respiration. Carbon dioxide is the main stimulus for respiration, which is activated by small elevations in the PCO₂. Hypoxia is a minor stimulus for respiration and is typically effective when the PO₂ is in the range of 50 to 55 mm Hg. Derangements in respiratory CO₂ excretion lead to alterations in the ratio of PCO₂ to bicarbonate in the serum and therefore alter systemic pH (recall the Henderson-Hasselbalch relationship, Eq. 4).

Respiratory Acidosis

Respiratory acidosis results from the primary retention of carbon dioxide; a variety of disorders that reduce ventilation can lead to respiratory acidosis. The common etiologies of respiratory acidosis seen in intensive care unit patients are listed in Table 42.8.

The increase in the plasma PCO₂ decreases the pH by formation of carbonic acid (Eq. 11). The principal compensatory defense mechanisms against respiratory acidosis are buffering and renal compensation. Recalling the Henderson-Hasselbalch relationship (Eq. 4), the pH of the blood is dependent on the relative concentrations of CO₂ and bicarbonate. Therefore, when there is an increase in PCO₂, the renal response to increase HCO₃^- is an action to normalize this relationship (Fig. 42.3B). In respiratory acidosis, the extracellular buffering capacity is severely limited because bicarbonate cannot buffer carbonic acid. Intracellular buffers—hemoglobin and other intracellular proteins—serve as the protection against acute rises in PCO₂. In circulating erythrocytes, the H⁺ that is produced as carbonic acid is formed from CO₂ that is buffered by hemoglobin; bicarbonate then leaves the cell in exchange for chloride. The buffering response to an elevation of CO₂ occurs within 10 to 15 minutes.

Renal compensation occurs in response to chronic respiratory acidosis. Hypercapnia stimulates secretion of protons in the distal nephron. Additionally, the urinary pH decreases and urinary ammonium excretion is increased, as is titratable acid excretion and excretion of chloride. The net effect is enhanced reabsorption of bicarbonate. The kidney's response to an acute increase in PCO₂ through compensation takes 3 to 4 days to reach completion (Table 42.2).

Aside from the compensatory mechanisms mentioned above, an increase in alveolar ventilation is ultimately required in order to eliminate excess CO₂ and therefore to re-establish equilibrium. If ventilation increases quickly during the acute period, the decrease in PCO₂ re-establishes equilibrium. However, following sustained hypercapnia that has elicited an appropriate renal response (i.e., a compensatory increase in serum bicarbonate), bicarbonaturia accompanies the return of the PCO₂ to normal. However, in order for this to occur, the chloride intake must be sufficient to replenish the deficit that developed during the renal compensation to the chronic respiratory acidosis, which induces a negative chloride balance. If chloride is deficient, the serum bicarbonate will remain persistently elevated, a phenomenon termed posthypercapnic metabolic alkalosis.

Table 42.8 Causes of respiratory acidosis

Airway obstruction · Foreign body, aspiration · Obstructive sleep apnea · Laryngospasm or bronchospasm Neuromuscular disorders of respiration Acute: Chronic: · Myasthenia gravis · Guillain-Barré syndrome · Botulism, tetanus, drugs · Hypokalemia, hypophosphatemia (respiratory muscle impairment) · Cervical spinal injury · Morbid obesity (aka Pickwickian syndrome) · Central sleep apnea Central respiratory depression · Drugs (opiates, sedatives), anesthetics · Oxygen administration in acute hypercapnia · Brain trauma or stroke Respirator disorders Acute: Chronic: · Severe pulmonary edema · Asthma or pneumonia · Acute respiratory distress syndrome · Chronic obstructive pulmonary disorder · Pulmonary fibrosis · Chronic pneumonitis

Clinical Presentation

Acute respiratory acidosis can produce headaches, confusion, irritability, anxiety, and insomnia, although the symptoms are difficult to separate from concomitant hypoxemia. Symptoms may progress to asterixis, delirium, and somnolence. The severity of the clinical presentation correlates more closely with the rapidity of the development of the disturbance rather than the absolute PCO₂ level.

Treatment

The treatment of respiratory acidosis is focused on alleviating the underlying disorder. In patients with acute respiratory acidosis and hypoxemia, supplemental oxygen is appropriate. However, to treat the hypercapnia, an increase in effective alveolar ventilation is necessary through either reversal of the underlying cause or, if necessary, mechanical ventilation. The administration of bicarbonate in respiratory acidosis when a coexisting metabolic acidosis is not present is potentially harmful. Bicarbonate in the setting of acute respiratory acidosis may precipitate acute pulmonary edema, metabolic alkalosis, and augmented carbon dioxide production, leading to increased PCO₂ in patients with inadequate respiratory reserve (45).

During chronic respiratory acidosis, renal compensation leads to a near-normalization of the arterial pH. In treating chronic respiratory acidosis, the objective is to ensure adequate oxygenation and, if possible, to increase alveolar ventilation. The administration of excessive oxygen and use of sedatives should be avoided because these treatments can depress the respiratory drive. Mechanical ventilation may be indicated when there is an acute exacerbation of chronic hypercapnia. If mechanical ventilation is used, the PCO₂ should be decreased gradually, avoiding precipitous drops, as rapid correction may cause severe alkalemia. Also, this may increase the cerebrospinal fluid pH, because carbon dioxide rapidly equilibrates across the blood–cerebrospinal fluid barrier. This complication can lead to serious neurologic problems, including seizures and death.

Special Scenario: Permissive Hypercapnia

It has been shown that ventilator strategies to reduce ventilator-associated lung injury (VALI) improve intensive care unit outcomes (46,47,48,49). This strategy is referred to as permissive hypercapnia and may reduce VALI through several mechanisms: by reducing stretch trauma and associated release of cytokines, and by preventing translocation of endotoxin and bacteria across the alveolar capillary barrier (50,51,52,53,54). It is not known for certain if respiratory acidosis per se has a beneficial effect, though there are recent data to suggest such an effect (55). Primary elevation of PCO₂ is also suggested to be deleterious on cardiac function (27,28), though this may be outweighed by protective effects of hypercapnia on lung injury (56). Further studies will be needed to delineate the specific roles of low tidal volume and respiratory acidosis with or without buffering in the management of acute lung injury.

Case Scenario #7: Respiratory Acidosis

A 56-year-old morbidly obese patient is admitted to the hospital with severe cellulitis of the right lower extremity that fails to respond to intravenous antibiotics. On hospital day 3, he undergoes a right below-knee amputation and, although recovering well, complains of severe pain postoperatively. His blood cultures drawn at admission are negative. His past medical history is significant for diabetes mellitus and chronic lower extremity ulceration that is felt to be secondary to venous stasis. Current medications include hydrochlorothiazide 25 mg PO daily, amlodipine 10 mg PO daily, enalapril 10 mg PO bid, metformin 1,000 mg PO bid, and a fentanyl patch 25 µg/hour. Two days following the operation he is found to have diminished mental status and the following laboratory data are obtained:

Serum
Sodium (mEq/L)	140
Potassium (mEq/L)	4.4
Chloride (mEq/L)	98
Bicarbonate (mEq/L)	34
BUN (mg/dL)	19
Creatinine (mg/dL)	1.2
Phosphorus (mg/dL)	4.3
Albumin (g/dL)	3.7
Glucose (mg/dL)	180
Arterial blood gas
pH	7.09
PO2 (mm Hg)	55
PCO2 (mm Hg)	110

· Which medication contributed most to the acid-base disorder prior to initiation of mechanical ventilation?

The laboratory values are consistent with acute respiratory acidosis, secondary to fentanyl, likely in the setting of a chronic respiratory acidosis, itself probably secondary to restrictive lung disease from obesity. The chronic respiratory acidosis is evidenced by the increased serum bicarbonate with a decreased pH. This is consistent with a history of obesity, which can lead to a restrictive pattern of lung disease characterized by chronic respiratory insufficiency. An acute respiratory acidosis is present because the expected degree of renal compensation is not present as it would be if the patient had a chronic elevation of the PCO₂ to levels of 110 mm Hg. The acute respiratory failure is most likely secondary to narcotic overdose, and thus fentanyl is the most likely causative agent.

Respiratory Alkalosis

Pathophysiology

Respiratory alkalosis is due to a primary increase in ventilation, which leads to a decrease in the PCO₂ and occurs commonly in the intensive care unit either as a treatment (e.g., for elevated intracranial pressure), as an iatrogenic complication of mechanical ventilation, or as part of a disease presentation (Table 42.9) (57). The lowered PCO₂ in turn reduces carbonic acid levels, which decreases systemic pH. The buffering system and, ultimately, renal compensation are the counterregulatory measures that are directed at maintaining plasma pH in this setting. In the setting of acute respiratory alkalosis, proteins, phosphates, and hemoglobin liberate H⁺. These liberated protons subsequently react with bicarbonate to form carbonic acid. At the level of the erythrocyte, a shift of chloride to the extracellular compartment ensues, as bicarbonate and cations enter in exchange for protons. The net effect of this buffering system reduces plasma pH and accounts for a 2 mEq/L decrease in the serum bicarbonate for every 10 mm Hg decrease in the PCO₂ that occurs in the acute setting (Table 42.2). Persistent respiratory alkalemia elicits the renal response, which leads to a net decrease in the secretion of H⁺. This renal compensation causes a decrease in the proximal reabsorption of bicarbonate and a decrease in the excretion of titratable acids and of ammonium. Recall that the excretion of one H⁺ in the kidney leads to the regeneration of a HCO₃^- molecule; therefore, these renal changes decrease renal production of HCO₃^-. This compensatory response is maximal 3 to 4 days following the onset of alkalemia and leads to further decrease in serum HCO₃^-.

Table 42.9 Causes of respiratory alkalosis

Hypoxia · High altitude · Congestive heart failure · Severe V/Q mismatch Lung diseases · Pulmonary fibrosis · Pulmonary edema · Pneumonia · Pulmonary embolism Drugs · Salicylates · Progesterone · Nicotine Direct stimulation of respiratory drive · Psychogenic hyperventilation · Cirrhosis · Gram-negative sepsis · Pregnancy (progesterone) · Excessive mechanical ventilation · Neurologic disorders (e.g., pontine tumors)

Clinical Presentation

Respiratory alkalosis may lead to a wide range of clinical manifestations ranging from alteration in consciousness, perioral paresthesias, and muscle spasms to cardiac arrhythmias. In addition, alkalemia also can affect metabolism of divalent ions. By stimulating glycolysis, alkalemia causes phosphate to shift from the extracellular space into the intracellular compartment as glucose-6-phosphate is formed. Additionally, the level of ionized calcium in the blood may also decrease due to increased binding of calcium to albumin.

Treatment

The underlying cause for respiratory alkalosis should be sought (Table 42.9). Cirrhosis can lead to respiratory alkalosis through impaired clearance from the circulation of progesterones and estrogens, similar to pregnancy (58). This is a commonly seen acid-base disorder in the intensive care unit. In psychogenic hyperventilation, rebreathing air using a bag increases the systemic PCO₂ and can treat alkalemia. Specific therapy, other than treatment of the underlying cause, is typically not necessary.

Case Scenario #8: Severe Acute Respiratory Alkalosis: A Potential Complication of Mechanical Ventilation

A 42-year-old patient with morbid obesity is admitted to the hospital with shortness of breath and fever. In the emergency room, he is started on intravenous antibiotics. Over the next 3 hours, he becomes severely short of breath and develops a diminished level of consciousness. He is intubated and placed on mechanical ventilation. His past medical history is significant for diabetes mellitus and hypertension. The social history is significant for one pack per day of smoking for 20 years. Current medications include amlodipine 5 mg PO daily, enalapril 5 mg PO bid, and hydrochlorothiazide 12.5 mg PO bid. Physical exam shows blood pressure of 156/80 mm Hg, pulse of 70 beats/minute, and temperature of 100.8°F. The patient is morbidly obese. The cardiovascular exam is normal. Lung exam reveals bilateral breath sounds with diffuse crackles on the right and egophony. Thirty minutes after mechanical ventilation laboratory studies are sent, with the following results:

Serum
Sodium (mEq/L)	140
Potassium (mEq/L)	4.4
Chloride (mEq/L)	97
Bicarbonate (mEq/L)	35
BUN (mg/dL)	15
Creatinine (mg/dL)	0.9
Phosphorus (mg/dL)	4.0
Albumin (g/dL)	3.9
Glucose (mg/dL)	146
Arterial blood gas
pH	7.66
PO2 (mm Hg)	340
PCO2 (mm Hg)	31

· What PCO₂ goal should be targeted in order to correct the acid-base disorder and attain a normal pH?

The respiratory rate should be decreased to maintain the pH at a level of about 55 mm Hg, as this will lead to a pH of approximately 7.4. When a patient with chronic respiratory acidosis and appropriate renal compensation is placed on mechanical ventilation, he or she is at risk of developing a posthypercapnic metabolic alkalosis, as occurred in this case. Quickly lowering the PCO₂ in a patient with an elevated bicarbonate can lead to a dangerous degree of alkalemia. This occurs because mechanical ventilation can remove PCO₂ from the blood quickly, thus increasing the pH precipitously. However, it takes time for the kidney to adapt to the change in blood pH. In time, the kidney can adapt by decreasing bicarbonate reabsorption, leading to loss of bicarbonate in the urine; but this does not happen quickly, usually requiring a minimum of 24 to 36 hours. The appropriate measure is to decrease the minute ventilation to allow the PCO₂ to rise to a level that would lead to a normal or slightly acidic pH. The target PCO₂ can be calculated by using the Henderson-Hasselbalch equation; however, this degree of precision is not always necessary. The minute ventilation can simply be decreased and titrated to achieve the desired pH.

Mixed Acid-Base Disorders

Mixed acid-base disorders are more difficult to diagnose than simple acid-base disorders. A good general rule is to keep in mind that in patients with a known primary acid-base disorder, a mixed disorder needs to be suspected if the pH is normal or if the apparent “compensation” has led to a pH that is beyond the normal. For example, if a patient with metabolic acidosis has a pH of 7.47, this indicates an accompanying respiratory alkalosis since a compensation would not lead to an alkaline pH if the primary disorder is an acidosis.

Metabolic and Respiratory Acidosis

In this mixed disorder, respiratory compensation is insufficient for the degree of decrease in bicarbonate. The most extreme example of this mixed-condition disorder occurs following cardiopulmonary arrest. In this setting, there is a decrease in bicarbonate levels—a metabolic acidosis secondary to lactic acidosis—and retention of carbon dioxide secondary to respiratory arrest. The pH in this setting is very low. Mixed metabolic and respiratory acidosis is also commonly seen in patients with a primary metabolic acidosis and concomitant lung disease. The lung disease impairs the ability of the patient to increase the ventilatory rate to appropriately decrease the PCO₂. Furthermore, this combination of disorders can manifest as a patient with a chronically elevated PCO₂ and an inability to increase the serum bicarbonate. This would indicate a chronic respiratory acidosis and possibly a metabolic acidosis due to a “normalized” serum bicarbonate in a setting in which the bicarbonate would be expected to be elevated. Finally, the presence of an anion gap, despite a normal serum bicarbonate level, should raise the index of suspicion for this combination of conditions.

Metabolic and Respiratory Alkalosis

Acidemia is better tolerated than is alkalemia. For example, a pH of 7.2 is well tolerated, whereas a pH of greater than 7.6 is associated with significant mortality. Mixed metabolic and respiratory alkalosis can lead to a significant elevation in the pH and is therefore very serious. This condition typically occurs in patients on mechanical ventilation. Often, mechanical ventilation, by mandating a minimum minute ventilation, will not allow the patient to elevate the PCO₂ significantly in response to alkalemia. Frequently, the metabolic alkalosis in this setting is due to diuretic use, administration of bicarbonate solutions, or massive transfusions with a citrate load.

Respiratory Alkalosis and Metabolic Acidosis

Most commonly, this combination is seen in Gram-negative sepsis, which can stimulate the respiratory drive—resulting in respiratory alkalosis—and also cause circulatory collapse with subsequent lactic acidosis. In the medical intensive care unit, salicylate intoxication classically leads to a mixed metabolic acidosis and respiratory alkalosis (23). Salicylates directly lead to an anion gap metabolic acidosis, and they also directly stimulate respiration.

Approach to Mixed Acid-base Disorders

There is no simple algorithm to use in the approach to a mixed acid-base disorder. The approach outlined in Figure 42.5 assumes that only a single disorder is present. Complex acid-base problems should be suspected when the values cannot be explained by a single disorder and its compensation. An example of this might be a patient with lactic acidosis and an alkaline pH.

Key Points

1. In patients with a known primary acid-base disturbance, a mixed acid-base disorder should always be suspected if the pH is normal or the “compensation” has surpassed the normal pH.

2. Mixed metabolic and respiratory acidosis occurs when the respiratory compensation is insufficient for the degree of decrease in bicarbonate.

3. The presence of an increased anion gap despite normal serum bicarbonate levels should raise suspicion for mixed metabolic acidosis and metabolic alkalosis.

4. Non–anion gap metabolic acidosis and anion gap metabolic acidosis can coexist.

5. Gram-negative sepsis is a common cause of respiratory alkalosis and metabolic acidosis.

6. Mixed metabolic acidosis and metabolic alkalosis occurs commonly in diabetic ketoacidosis.

Case Scenario #9: Mixed Acid-base Disorder: Metabolic Acidosis and Respiratory Acidosis

A 64-year-old is admitted to the intensive care unit with pneumonia and septic shock. The patient states that he has had increasing shortness of breath and fever over the past 4 days. His past medical history is significant for hypertension. Surgical history is significant for a previous cholecystectomy. Medications include amlodipine and hydrochlorothiazide. Physical exam shows a blood pressure of 85/50 mm Hg, pulse of 110 beats/minute, respiratory rate of 22 breaths/minute, and temperature of 101.8°F. The cardiovascular examination is significant for a 2/6 systolic murmur and there are crackles over his entire right lung field. There is trace pedal edema. Chemistry values on admission are given below:

Serum
Sodium (mEq/L)	135
Potassium (mEq/L)	4.8
Chloride (mEq/L)	103
Bicarbonate (mEq/L)	10
BUN (mg/dL)	22
Creatinine (mg/dL)	1.4
Phosphorus (mg/dL)	2.8
Albumin (g/dL)	3.8
Glucose (mg/dL)	115
Arterial blood gas
pH	6.95
PO2 (mm Hg)	51
PCO2 (mm Hg)	48

· What acid-base disorder(s) is (are) present in this patient?

The acid-base disorder is a mixed anion gap metabolic acidosis with a respiratory acidosis. The decrease in bicarbonate accompanied by an elevated anion gap is consistent with a primary metabolic acidosis. The expected PCO₂ using the Winter formula = 10(1.5) + 4 = 19. The PCO₂ is significantly elevated above this level, and thus a respiratory acidosis is present. The PCO₂ is much higher than would be expected based on the degree of acidemia, and thus a respiratory acidosis, secondary to inadequate ventilation from pneumonia, is present.

References

1. Jorgensen K, Astrup P. Standard bicarbonate, its clinical significance, and a new method for its determination. Scand J Clin Lab Invest. 1957;9(2):122–132.

2. Batlle DC, Hizon M, Cohen E, et al. The use of the urinary anion gap in the diagnosis of hyperchloremic metabolic acidosis. N Engl J Med. 1988;318(10):594–599.

3. Gabow PA. Disorders associated with an altered anion gap. Kidney Int. 1985;27(2):472–483.

4. Winter SD, Pearson JR, Gabow PA, et al. The fall of the serum anion gap. Arch Intern Med. 1990;150(2):311–313.

5. Sadjadi SA. A new range for the anion gap. Ann Intern Med. 1995;123(10):807.

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.

7. Fencl V, Jabor A, Kazda A, et al. Diagnosis of metabolic acid-base disturbances in critically ill patients. Am J Respir Crit Care Med. 2000;162(6):2246–2251.

8. Moe OW, Fuster D. Clinical acid-base pathophysiology: disorders of plasma anion gap. Best Pract Res Clin Endocrinol Metab. 2003;17(4):559–574.

9. Hassan H, Joh JH, Bacon BR, et al. Evaluation of serum anion gap in patients with liver cirrhosis of diverse etiologies. Mt Sinai J Med. 2004;71(4):281–284.

10. Schnur MJ, Appel GB, Karp G, et al. The anion gap in asymptomatic plasma cell dyscrasias. Ann Intern Med. 1977;86(3):304–305.

11. Murray T, Long W, Narins RG. Multiple myeloma and the anion gap. N Engl J Med. 1975;292(11):574–575.

12. Madias NE, Ayus JC, Adrogue HJ. Increased anion gap in metabolic alkalosis: the role of plasma-protein equivalency. N Engl J Med. 1979;300(25):1421–1423.

13. Salem MM, Mujais SK. Gaps in the anion gap. Arch Intern Med. 1992;152(8):1625–1629.

14. Story DA, Kellum JA. New aspects of acid-base balance in intensive care. Curr Opin Anaesthesiol. 2004;17(2):119–123.

15. Corey HE. Stewart and beyond: new models of acid-base balance. Kidney Int. 2003;64(3):777–787.

16. Wooten EW. Science review: quantitative acid-base physiology using the Stewart model. Crit Care. 2004;8(6):448–452.

17. Story DA. Bench-to-bedside review: a brief history of clinical acid-base. Crit Care. 2004;8(4):253–258.

18. Kellum JA, Kramer DJ, Pinsky MR. Strong ion gap: a methodology for exploring unexplained anions. J Crit Care. 1995;10(2):51–55.

19. Feldman M, Soni N, Dickson B. Influence of hypoalbuminemia or hyperalbuminemia on the serum anion gap. J Lab Clin Med. 2005;146(6):317–320.

20. Pierce NF, Fedson DS, Brigham KL, et al. The ventilatory response to acute base deficit in humans. Time course during development and correction of metabolic acidosis. Ann Intern Med. 1970;72(5):633–640.

21. Adrogue HJ, Wilson H, Boyd AE 3rd, et al. Plasma acid-base patterns in diabetic ketoacidosis. N Engl J Med. 1982;307(26):1603–1610.

22. Judge BS. Differentiating the causes of metabolic acidosis in the poisoned patient. Clin Lab Med. 2006;26(1):31–48, vii.

23. Krause DS, Wolf BA, Shaw LM. Acute aspirin overdose: mechanisms of toxicity. Ther Drug Monit. 1992;14(6):441–451.

24. Kurtzman NA. Disorders of distal acidification. Kidney Int. 1990;38(4):720–727.

25. Narins RG, Goldberg M. Renal tubular acidosis: pathophysiology, diagnosis and treatment. Dis Mon. 1977;23(6):1–66.

26. DeFronzo RA. Hyperkalemia and hyporeninemic hypoaldosteronism. Kidney Int. 1980;17(1):118–134.

27. Arieff AI, Leach W, Park R, et al. Systemic effects of NaHCO3 in experimental lactic acidosis in dogs. Am J Physiol. 1982;242(6):F586–591.

28. Graf H, Leach W, Arieff AI. Evidence for a detrimental effect of bicarbonate therapy in hypoxic lactic acidosis. Science. 1985;227(4688):754–756.

29. Jeffrey FM, Malloy CR, Radda GK. Influence of intracellular acidosis on contractile function in the working rat heart. Am J Physiol. 1987;253(6 Pt 2):H1499–1505.

30. Ayus JC, Krothapalli RK. Effect of bicarbonate administration on cardiac function. Am J Med. 1989;87(1):5–6.

31. Kalkut G. Antiretroviral therapy: an update for the non-AIDS specialist. Curr Opin Oncol. 2005;17(5):479–484.

32. Wohl DA, McComsey G, Tebas P, et al. Current concepts in the diagnosis and management of metabolic complications of HIV infection and its therapy. Clin Infect Dis. 2006;43(5):645–653.

33. Moyle GJ, Datta D, Mandalia S, et al. Hyperlactataemia and lactic acidosis during antiretroviral therapy: relevance, reproducibility and possible risk factors. Aids. 2002;16(10):1341–1349.

34. Tebb Z, Tobias JD. New anticonvulsants–new adverse effects. South Med J. 2006;99(4):375–379.

35. Guo PY, Storsley LJ, Finkle SN. Severe lactic acidosis treated with prolonged hemodialysis: recovery after massive overdoses of metformin. Semin Dial. 2006;19(1):80–83.

36. Parke TJ, Stevens JE, Rice AS, et al. Metabolic acidosis and fatal myocardial failure after propofol infusion in children: five case reports. BMJ. 1992;305(6854):613–616.

37. Cremer OL, Moons KG, Bouman EA, et al. Long-term propofol infusion and cardiac failure in adult head-injured patients. Lancet. 2001;357(9250):117–118.

38. Perrier ND, Baerga-Varela Y, Murray MJ. Death related to propofol use in an adult patient. Crit Care Med. 2000;28(8):3071–3074.

39. Vasile B, Rasulo F, Candiani A, et al. The pathophysiology of propofol infusion syndrome: a simple name for a complex syndrome. Intensive Care Med. 2003;29(9):1417–1425.

40. Corbett SM, Moore J, Rebuck JA, et al. Survival of propofol infusion syndrome in a head-injured patient. Crit Care Med. 2006;34(9):2479–2483.

41. Wolf A, Weir P, Segar P, et al. Impaired fatty acid oxidation in propofol infusion syndrome. Lancet. 2001;357(9256):606–607.

42. Cray SH, Robinson BH, Cox PN. Lactic acidemia and bradyarrhythmia in a child sedated with propofol. Crit Care Med. 1998;26(12):2087–2092.

43. Luke RG, Galla JH. Chloride-depletion alkalosis with a normal extracellular fluid volume. Am J Physiol. 1983;245(4):F419–424.

44. Galla JH, Bonduris DN, Luke RG. Effects of chloride and extracellular fluid volume on bicarbonate reabsorption along the nephron in metabolic alkalosis in the rat. Reassessment of the classical hypothesis of the pathogenesis of metabolic alkalosis. J Clin Invest. 1987;80(1):41–50.

45. Ayus JC, Olivero JJ, Adrogue HJ. Alkalemia associated with renal failure. Correction by hemodialysis with low-bicarbonate dialysate. Arch Intern Med. 1980;140(4):513–515.

46. Amato MB, Barbas CS, Medeiros DM, et al. Effect of a protective-ventilation strategy on mortality in the acute respiratory distress syndrome. N Engl J Med. 1998;338(6):347–354.

47. Hickling KG, Walsh J, Henderson S, et al. Low mortality rate in adult respiratory distress syndrome using low-volume, pressure-limited ventilation with permissive hypercapnia: a prospective study. Crit Care Med. 1994;22(10):1568–1578.

48. Brower RG, Rubenfeld GD. Lung-protective ventilation strategies in acute lung injury. Crit Care Med. 2003;31(4 Suppl):S312–316.

49. Laffey JG, O'Croinin D, McLoughlin P, et al. Permissive hypercapnia–role in protective lung ventilatory strategies. Intensive Care Med. 2004;30(3):347–356.

50. Maggiore SM, Jonson B, Richard JC, et al. Alveolar derecruitment at decremental positive end-expiratory pressure levels in acute lung injury: comparison with the lower inflection point, oxygenation, and compliance. Am J Respir Crit Care Med. 2001;164(5):795–801.

51. Boussarsar M, Thierry G, Jaber S, et al. Relationship between ventilatory settings and barotrauma in the acute respiratory distress syndrome. Intensive Care Med. 2002;28(4):406–413.

52. Slutsky AS, Tremblay LN. Multiple system organ failure. Is mechanical ventilation a contributing factor? Am J Respir Crit Care Med. 1998;157(6 Pt 1):1721–1725.

53. Tremblay L, Valenza F, Ribeiro SP, et al. Injurious ventilatory strategies increase cytokines and c-fos m-RNA expression in an isolated rat lung model. J Clin Invest. 1997;99(5):944–952.

54. Murphy DB, Cregg N, Tremblay L, et al. Adverse ventilatory strategy causes pulmonary-to-systemic translocation of endotoxin. Am J Respir Crit Care Med. 2000;162(1):27–33.

55. Kregenow DA, Rubenfeld GD, Hudson LD, et al. Hypercapnic acidosis and mortality in acute lung injury. Crit Care Med. 2006;34(1):1–7.

56. Sinclair SE, Kregenow DA, Lamm WJ, et al. Hypercapnic acidosis is protective in an in vivo model of ventilator-induced lung injury. Am J Respir Crit Care Med. 2002;166(3):403–408.

57. Laffey JG, Kavanagh BP. Hypocapnia. N Engl J Med. 2002;347(1):43–53.

58. Lustik SJ, Chhibber AK, Kolano JW, et al. The hyperventilation of cirrhosis: progesterone and estradiol effects. Hepatology. 1997;25(1):55–58.