Chapter 28: Acid-Base Disorders

Recognition of acid-base disturbances through interpretation of arterial blood gases is of fundamental importance to the day-to-day clinical practice of critical care. Knowledge of the underlying basic chemical relationships and established nomenclature is prerequisite to this understanding. Several different paradigms of acid-base relationships have been proposed and are in current clinical use for framing the results of blood gas and related laboratory measurements. These paradigms include the base excess method, the Stewart or physicochemical method, and the bicarbonate-pH-Pco₂ method, and they all basically arrive at common endpoints of interpretation.^1,2 The bicarbonate-pH-Pco₂ method, also known as the physiological method, will be described herein. It is based on the carbonic acid-bicarbonate buffer system, is well established, has an empiric basis, and is in wide clinical use. Starting with a review of acid-base terminology and basic chemical relationships, this chapter will describe the basis of the physiological method for arriving at a diagnosis of simple and mixed acid-base disorders, and show how readily available laboratory testing, in conjunction with clinical information from the history and physical examination, can narrow the differential diagnosis with the goal of pin-pointing the underlying causative process or disease.

One definition for an acid is a chemical compound capable of donating a hydrogen ion (ie, a proton). The degree of acidity of a solution can thus be quantified as the hydrogen ion activity, which is closely related to the molar concentration of hydrogen ions. Normal hydrogen ion concentration of plasma averages about 40 nmol/L, which is considerably less than other plasma cations commonly assayed in the clinical setting. For example, plasma sodium concentration normally averages about 140 mmol/L, which is 3.5 million-fold higher than normal H⁺ concentration. To avoid using small numbers and to allow simple expression of H⁺ activity or concentration over a wide range of values, the pH scale was developed. The relationship between pH and hydrogen ion concentration is defined by

where the square brackets here signify molar concentration units.

A buffer is a solution containing a weak acid (HA) and its conjugate base (A^-), which can coexist in an equilibrated reversible chemical reaction represented generically as

The molar proportions of the substances represented in this chemical equation assume definite proportional relationships to one another at equilibrium, dependent on the chemical nature of the specific acid, temperature, and certain other variables. These proportions are expressible numerically by the modified equilibrium constant (K′) as

The prime symbol is added to signify that molar concentrations are employed rather than thermodynamic activities.³ Buffers resist changes in pH when additional acid or base is introduced into the solution because the added acid reacts with the conjugate base, while added base reacts with the existing acid or the acid's dissociated protons. Bodily fluids contain multiple buffers, but the proportions of their acid form to their conjugate base form are all related by way of the shared hydrogen ion concentration of the fluid as

Thus, only one buffer pair needs to be considered to describe the system in terms of [H⁺] or pH. Carbonic acid and its conjugate base (bicarbonate) are selected as the representative buffer pair because of the ability of carbonic acid to dissociate to form carbon dioxide, the body's ability to excrete large amounts of carbon dioxide by way of the lungs, and ready availability of suitable assay methods.

In the presence of carbonic anhydrase, the reversible dissociation of carbonic acid to carbon dioxide occurs rapidly and in accordance with the principle of chemical mass action as

Following the general form of equation (28-2), carbonic acid can also form ionic dissociation products, a hydrogen cation and bicarbonate anion, as

Combining equations (28-5) and (28-6) yields²

Assuming equilibrium conditions, and substituting the carbonic acid dissociation reaction in equation (28-6) into the general form given in equation (28-3), yields³

K′ for this reaction can be empirically demonstrated to be 7.9 × 10^–7. Applying the negative logarithm function, and using analogous symbolism as employed by equation (28-1), yields a more convenient form for the constant, pK′, equal to approximately 6.1.⁴ The stoichiometric relationship shown in equation (28-5) allows substitution of CO₂ for carbonic acid as

Molar concentration units of carbon dioxide are related to clinically familiar gas tension units by way of the gas's solubility coefficient, α (0.03 mmol/L per torr at 37°C), as

Substituting equation (28-10) into equation (28-9), supplying numerical approximations for the equilibrium and solubility constants, and using units of nanomoles per liter for [H⁺], millimoles per liter for [HCO₃^–], and torr (equivalent to millimeters of mercury) for Pco₂, yields

which is known as the Henderson equation. Taking the negative logarithm of both sides of equation (28-9), and performing substitutions allowed by equations (28-1) and (28-10), yields^2,4

which is known as the Henderson–Hasselbalch equation, basically a logarithmic form of the Henderson equation. Note that the K′ term has been replaced by the symbol pK′, the negative logarithm of the modified equilibrium constant. Although equation (28-12) is based on equilibrium conditions, the relationship between the three acid-base variables pH, bicarbonate concentration, and carbon dioxide tension empirically prevail, even in unstable critically ill patients.⁴ This relationship also provides the basis for determining bicarbonate concentration in virtually all commercial blood gas analyzers, which utilize a hydrogen ion-selective glass electrode to determine pH and a Severinghaus electrode to determine Pco₂, and then solve equation (28-12) for bicarbonate concentration. Either the Henderson equation or the Henderson–Hasselbalch equation can also be used in the clinical setting to check the internal consistency of the three acid-base variables obtained by blood gas analysis. This is accomplished by selecting any two of the reported variables, calculating the third variable, and comparing the calculated value to that reported. A significant discrepancy (after accounting for rounding to significant digits) signifies that one or more of the reported values is faulty, usually due to transcription error.

Multiphasic plasma or serum electrolyte profiles usually include measurements of sodium, potassium, chloride, and total carbon dioxide content (tCO₂). The latter analyte is often used interchangeably with bicarbonate concentration in clinical parlance. However, tCO₂ actually represents the combined molar concentrations of CO₂ and various acid-labile precursor forms of carbon dioxide. These include bicarbonate and several other chemical entities. Thus,

where R - NH - COO^- represents labile carbamino compounds formed by combination of carbon dioxide with amine moieties of plasma proteins by

Quantitatively, however, the main constituents of tCO₂ are bicarbonate and dissolved CO₂ gas. Thus, equation (28-13) can be closely approximated by eliminating the negligible terms, and substituting from equation (28-10), as³

where total CO₂ content and bicarbonate concentration are given in millimoles per liter and Pco₂ is given in torr. Substituting average normal values for bicarbonate concentration (24.0 mmol/L) and arterial Pco₂ (Paco₂), 40 torr, yields a tCO₂ level of 25.2 mmol/L, which is only marginally higher than the bicarbonate concentration. However, in patients with marked hypercapnia, more significant discrepancies may be seen. For example, at a Paco₂ of 100 torr and bicarbonate concentration of 28.0 mmol/L, tCO₂ would equal 31.0 mmol/L.

A simple acid-base disorder is one in which there is either acidosis or alkalosis (Table 28–1), but not both simultaneously, and this primary pathologic process is either of metabolic or respiratory origin, but not both. If sufficiently severe, a simple acid-base disturbance will always result in an abnormality in blood pH, that is, either alkalemia or acidemia. Thus, there are four cardinal acid-base disorders: metabolic acidosis, metabolic alkalosis, respiratory acidosis, and respiratory alkalosis. The patterns of pH, Paco₂, and bicarbonate concentration representing these cardinal disturbances, uncomplicated by any superimposed acid-base disorder, have been observed in animal models in which the cardinal disturbance is imposed by an experimental procedure (eg, infusing an acid, such as HCl or lactic acid intravascularly, or manipulating minute ventilation in mechanically ventilated animals). These patterns have also been observed clinically in patients by independently surmising the particular acid-base process present and then examining the associated arterial blood gas assay results. Limited studies have also been performed in healthy human volunteers. The next four sections explain the patterns observed in these simple cardinal disorders, delineate the associated quantitative relationship between blood gas variables, list the differential diagnosis (ie, the underlying causes) of each cardinal disorder, and provide some simple methods for narrowing the differential diagnosis.

Metabolic Acidosis

From equation (28-7), it can be seen that loss of bicarbonate from the body, for example, from diarrhea, will result in formation of excessive hydrogen ion concentration by the principle of chemical mass action. On the other hand, addition of acid, either exogenous or generated by pathophysiologic metabolic processes, can directly increase hydrogen ion concentration and therefore lower pH. Also, equation (28-7) shows that addition of acid (H⁺) to body fluids generates carbon dioxide above and beyond the amount would be produced by normal physiologic mechanisms. Blood pH is expected to decrease from the added hydrogen ions that have not reacted with bicarbonate, and bicarbonate concentration will decrease from the hydrogen ions that combine with bicarbonate to form carbon dioxide. To prevent hypercapnia, the body has a built-in reflexive mechanism for dealing with this increased CO₂ generation, namely, hyperventilation mediated through the brainstem at the level of the medulla. The afferent limb of this physiologic reflex reacts to the fall in pH to neurogenically effect hyperventilation. The increase in ventilation does not simply suffice to maintain baseline Paco₂, but rather it lowers Paco₂ below the normal range to a degree that is proportional to the extent of the metabolic acidosis. Empiric analysis of this proportionality in humans has been found to be linear, and can be represented by⁵

where HCO₃^– is the patient's observed bicarbonate concentration (in millimoles per liter), and expected Paco₂ is the arterial carbon dioxide gas tension (in torr) that occurs statistically (ie, on average) in uncomplicated metabolic acidosis of a severity level corresponding to the observed bicarbonate concentration in the patient at hand.

Equation (28-16) follows the general form of a linear equation: y = mx + b, where m represents the slope of the linear equation when graphed on Cartesian coordinates with the X-axis representing HCO₃^– concentration and the Y-axis representing Paco₂, and b represents the point of interception of the line along the Y-axis. The slope, given above by the factor 1.5 in equation (28-16), has been shown in some studies to be closer to 1.2.⁶ The term ±2 is appended to represent the 95% confidence interval derived from empiric data. If simple (ie, uncomplicated by another primary acid-base disorder) metabolic acidosis is present, substituting the patient's bicarbonate concentration into equation (28-16) and finding the patient's Paco₂ to be within the expected range derived from solving the equation, would signify that the patient's arterial blood gas findings are consistent with metabolic acidosis. A Paco₂ value lying outside this range implies that metabolic acidosis either is not present, is incompletely developed, or is present in combination with some other cardinal acid-base disturbance. The latter situation would constitute a mixed acid-base disturbance.

Table 28–2 provides a differential diagnosis for metabolic acidosis. A classic method for narrowing this differential diagnosis is to use the results of a basic multiphasic serum (or plasma) electrolyte panel to arrive at the anion gap by^7,8,9

where each analyte concentration is expressed in units of millimoles (or milliequivalents) per liter. The normal range for the serum anion gap varies among clinical chemistry laboratories, and has declined somewhat following the adoption of contemporary assay methods, but for some laboratories, it is 12 ± 4 mEq/L. Values exceeding the upper normal limit imply that an anion, other than chloride or bicarbonate, is present at pathological concentrations. The cause of abnormally high anion gap values, particularly if the degree of elevation is not subtle, often can be confirmed, in whole or in part, by measurement of the culpable anion; for example, lactate if the cause is lactic acidosis, or β-hydroxybutyrate if the cause is ketoacidosis. Assays for some organic anions that can increase the anion gap, for example, glycolate in ethylene glycol poisoning as well as some inorganic anions, for example, sulfate accumulation in chronic renal failure, are not commonly available; however, information from the patient's medical history, physical examination, and ancillary laboratory assays can often corroborate the specific diagnosis in these cases.

If the anion gap is normal, the cause of the metabolic acidosis is basically either loss of bicarbonate from the body or an etiology that does not involve excessive production of an organic acid. In cases of metabolic acidosis where the serum anion gap is not elevated, clinical information coupled with other simple laboratory tests can often facilitate delineation of the cause.^10,11 In cases of mild metabolic acidosis due to organic anion accumulation, the corresponding mild perturbation in the anion gap may not be sufficient to increase the gap above the upper normal limit. Similarly, a low baseline serum anion gap, as can occur in marked hypoalbuminemia, can have the same effect.⁸ The latter phenomenon can be taken into account by measuring serum albumin and recognizing that for every 1 g/dL decrease in serum albumin concentration, the serum anion gap decreases by about 2.4 mEq/L, on average.¹²

The serum osmole gap, often used as a screening test when methanol or ethylene glycol intoxication is suspected, is another simple derived laboratory-based expression that can be helpful in the differential diagnosis of metabolic acidosis when either of these poisonings is suspected.¹³ The osmole gap can be elevated after methanol or ethylene glycol ingestion even before the metabolic acidosis develops. As methanol is metabolized to formate, or ethylene glycol is metabolized to glycolate, both of which are acid anions, the serum osmole gap declines, the serum anion gap rises, and metabolic acidosis evolves.¹⁴ The osmole gap is the difference between serum osmolality determined by freezing point depression osmometry (Osm_fp) and serum osmolality estimated from routinely available clinical chemistry testing (Osm_est), expressed as

Osm_est is calculated from simultaneous measurements of the main osmotically active constituents of normal plasma: sodium, urea nitrogen (BUN), and glucose, along with ethanol, if applicable, by:

where Na⁺ is expressed in millimoles per liter and the other serum assay results are expressed in milligrams per deciliter. The sodium concentration is doubled to account for osmotically active attendant anions that balance the positive charges of the sodium ions. The divisors in the other terms are derived from their respective molecular weights and convert their concentration units to millimoles (equivalent to milliosmoles) per liter. An elevated osmole gap (>15 mOsm/kg H₂O or so) suggests the presence of occult osmoles, such as methanol or ethylene glycol molecules present in the sample and contributing to total osmolality.¹³ Other potential causes of an elevated serum osmole gap include ethanol intoxication (if the ethanol term is deleted from equation [28-19]), ketoacidosis, and severe degrees of circulatory shock.^9,13

Metabolic Alkalosis

Addition of bicarbonate or loss of hydrogen ions from the body elevates extracellular bicarbonate concentration, decreases hydrogen ion concentration, and correspondingly increases pH. As in metabolic acidosis, there is normally an associated reflexive change in pulmonary minute ventilation, but in this case manifesting as hypoventilation and resulting in CO₂ retention. The resulting hypercapnia does not represent respiratory dysfunction, but is an expected physiologic response to the metabolic alkalosis. The proportional change in ventilation, as manifested by Paco₂, is related to the degree of metabolic alkalosis as expressed by the plasma bicarbonate concentration. One empirically derived equation describing this relationship is

The slope, given above as 0.9, has been shown in some studies to be closer to 0.7.⁶ Greater variability in this relationship has been observed compared to that seen in metabolic acidosis. Nevertheless, arterial blood gas values that approximate this relationship are consistent with simple metabolic alkalosis.

Once the diagnosis of metabolic alkalosis is established, the range of possible mechanisms or etiologies is considered (Table 28–3). As with metabolic acidosis, the applicable etiology is often apparent from information available through the medical history and physical examination.^11,15 Gastric fluid loss, diuretic use, and extracellular volume contraction are among the most common causes. Excess alkali intake (< 10 mEq/kg) is excreted rapidly in most patients with normal renal function. However, exogenous alkali administration can lead to alkalosis in patients with compromised renal function, low chloride intake, or mineralocorticoid excess.¹⁶ Sources of exogenous alkali include oral or parenteral bicarbonate (including oral baking soda or overtreatment with intravenous NaHCO₃), acetate (eg, in parenteral nutrition formulas or dialysate solutions), citrate (including secondary to multiple blood transfusions), lactate (including Ringer's lactate solutions), and gluconate salts (found, eg, in certain proprietary intravenous balanced salt solutions), or CaCO₃ (eg, in oral calcium supplements).

In cases when the etiology is not straightforward, a simple laboratory test, urine chloride measurement, can sometimes be helpful. Low urine chloride concentration (< 20 mmol/L), commonly points to gastric fluid losses (by vomiting or gastric suction), recently discontinued diuretic therapy, or a posthypercapnic state. The latter occurs commonly in patients with chronic carbon dioxide retention due to chronic obstructive pulmonary disease who are placed on mechanical ventilation with resultant rapid lowering of Paco₂. High urine chloride concentration (> 20 mmol/L), on the other hand, points to a renal mechanism that involves chloride wasting with retention of the alternative anion bicarbonate. Common causes of chloride wasting in the intensive care unit setting are active diuretic administration and corticosteroid therapy. The urine chloride concentration also has therapeutic implications. Low urine chloride concentrations in the face of metabolic alkalosis are often readily corrected by simply supplying chloride in the form of ample sodium chloride-containing intravenous fluids. Patients with high urine chloride levels, for example, mediated by excessive mineralocorticoid activity, are resistant to therapy with chloride-containing parenteral fluid administration because the administrated chloride load is excreted without correction of the alkalosis. Note that some patients with chloride-resistant forms of metabolic alkalosis may not manifest high urine chloride levels if they have been deprived of dietary or parenteral chloride.

Respiratory Acidosis

Acute hypoventilation in an otherwise normal subject results in decreased elimination of carbon dioxide, with a resultant rise in Paco₂. Accompanying this rise is an immediate concomitant rise in the plasma bicarbonate level. This rapid rise in bicarbonate concentration is not mediated by renal mechanisms, but rather by the principle of chemical mass action affecting equation (28-7), shifting accumulated CO₂ leftward to form bicarbonate. Controlled and observational studies have demonstrated that the average rise in bicarbonate concentration (here measured in millimoles per liter) is approximately one-tenth the increase in arterial CO₂ tension (measured in torr), that is,^2,6,17

When a hypercapnic patient's arterial blood gas values (Paco₂ tension and bicarbonate concentration) conform to equation (28-21) by comparison to known or assumed baseline (prehypercapnic) values, the results are consistent with simple acute respiratory acidosis.

The kidney normally reclaims bicarbonate filtered at the glomerulus by secreting hydrogen ions into the renal tubular lumen. The reaction between bicarbonate and hydrogen ions forms CO₂ by equation (28-7), and the CO₂ is readily absorbed from the tubules into the bloodstream where equation (28-7) then can operate in reverse, thus effectively resulting in bicarbonate reabsorption.² In sustained hypercapnia, the kidney normally responds by gradually increasing tubular hydrogen ion secretion, thereby enhancing bicarbonate reabsorption, adding to the bicarbonate-elevation effect given by equation (28-21). This renal effect on plasma bicarbonate concentration takes several days to fully evolve and is proportional to the degree of prevailing hypercapnia. Although the relationship is not precisely linear, the combined mass action and renal effects of chronic hypercapnia on bicarbonate concentration can be approximated by^2,6:

When the patient's arterial blood gas values yield a plasma bicarbonate concentration that approximately conforms to that given by equation (28-22), the results are consistent with simple chronic respiratory acidosis. Patients having a bicarbonate concentration intermediate between the values obtained using equations (28-21) and (28-22) either have an intermediate degree of chronicity with respect to their hypercapnia, or have a mixed acid-base disturbance (see the final section).

Etiologies of hypercapnia can be divided into primary pulmonary derangements, primary neurological abnormalities, and metabolically mediated causes, with the latter potentially including drug or toxin-mediated causes (Table 28–4). The patient's medical history, physical examination, chest imaging studies and, where necessary, other investigations will assist in narrowing the differential diagnosis and arriving at the reason for respiratory acidosis.

Respiratory Alkalosis

Acute hyperventilation in an otherwise normal subject results in increased elimination of carbon dioxide and therefore a fall in Paco₂. Accompanying this fall is an immediate concomitant decrease in plasma bicarbonate concentration. This rapid fall in bicarbonate is not mediated by renal mechanisms, but rather by the principle of chemical mass action affecting equation (28-7), causing a rightward shift in the reaction sequence to partially fill the void left by the expired CO₂. Human observational studies have shown that the fall in bicarbonate concentration (measured in millimoles per liter) is approximately two-tenths the decrease in arterial CO₂ tension (measured in torr), that is,⁶

Arterial blood gas values showing hypocapnia and decreased plasma bicarbonate concentrations that approximate those given by equation (28-23) are consistent with simple acute respiratory alkalosis.

As described earlier, the kidney normally regulates bicarbonate reabsorption by adjusting hydrogen ion secretion into the renal tubular lumen. Sustained hypocapnia downregulates this hydrogen ion secretion, resulting in less reclaimed bicarbonate and thus enhancing bicarbonate excretion, adding to the bicarbonate lowering effect accounted for by equation (28-23). This added renal effect on bicarbonate concentration takes several days to fully evolve and is proportional to the degree of prevailing hypocapnia, as given by^6,18

However, this proportionality, representing the combined mass action and renal bicarbonate wasting effects of hypocapnia on plasma bicarbonate concentration, eventually reaches a limit such that further degrees of hypocapnia have little additional effect on the plasma bicarbonate level. Thus, plasma HCO₃^– concentrations much less than 17 mmol/L are not commonly observed in simple chronic respiratory alkalosis. Within this limit, arterial blood gas results that show hypocapnia and plasma bicarbonate concentrations that conform to that given by equation (28-24) are consistent with simple chronic respiratory alkalosis. Patients having a bicarbonate concentration intermediate between the values obtained by equations (28-23) and (28-24) either have an intermediate degree of chronicity to their hypocapnia, or have a mixed acid-base disturbance (see the following section).

Etiologies of hypocapnia can be classified according to whether the primary disorder is pulmonary, neurological, metabolic, or a drug or toxin-mediated cause (Table 28–5). The patient's medical history, physical examination, chest imaging studies and, where necessary, other investigations will assist in narrowing the differential diagnosis and arriving at the reason for respiratory acidosis.

Whereas simple acid-base disturbances involve a single primary underlying causative mechanism, either metabolic or respiratory, mixed acid-base disorders involve more than one cardinal acid-base derangement simultaneously. Co-occurrence of two or three of the cardinal acid-base disorders is not uncommon in critically ill patients and frequently complicates interpretation of arterial blood gas determinations in the intensive care unit setting.

The data used to derive the empiric formulas given for the cardinal acid-base disorder in the preceding sections have also been used to generate statistical confidence limits for each disorder. These confidence bands are approximated graphically in Figure 28–1, a plot of Paco₂ versus bicarbonate concentration.^2,5,6,17,19 According to equation (28-12), a given Paco₂ and bicarbonate concentration defines a specific pH; therefore, lines of equal pH can be superimposed on the graph (the diagonal lines in Figure 28–1). The elliptical area near the center of the graph represents the 95% confidence area for the combined normal ranges for Paco₂ and bicarbonate concentration. The normal range for arterial blood pH (7.36-7.44) can be indirectly appreciated on the graph by interpolating between the diagonal lines. The dark bands represent confidence areas defining patterns of blood gas values that have been observed when the depicted cardinal acid-base disorder is present and is not complicated by another cardinal acid-base disturbance. For the respiratory disorders, both acute and chronic confidence bands are given. The overlap in the confidence bands near the center of the graph (at the outskirts of the normal confidence ellipse) illustrates that mild blood gas abnormalities can present greater diagnostic uncertainty compared to more severe abnormalities. Figure 28–1 also highlights the quantitative relationship between the three blood gas variables and shows that merely examining the directional changes in one or two of the variables, without considering the quantitative relationship between them, may not allow accurate identification of the cardinal disturbance. For example, finding that a patient has a low bicarbonate concentration does not allow one to distinguish whether the patient has metabolic acidosis or respiratory alkalosis. Similarly, knowing that a patient has a low bicarbonate level and low Paco₂ does not discriminate between these two disturbances. Quantifying the relationship between the acid-base variables, however, often identifies which disorder is present.

When one of the cardinal respiratory disturbances is present, and the time course of the disorder is intermediate between an acute and a fully developed chronic disturbance, the blood gas values will lie between the acute and chronic respiratory bands. Analogously, when two cardinal disturbances are present simultaneously (eg, metabolic acidosis and respiratory alkalosis), the range of possible arterial blood gas values can lie anywhere between the confidence areas given for the two cardinal disorders in Figure 28–1. Thus, abnormal blood gas findings that do not fall within any of the confidence bands imply that a mixed disturbance is present. If there are two cardinal disorders present, the point representing the blood gas values will usually lie between the two bands representing the disorders.

Note that although blood gas results lying within a confidence band are consistent with the disorder represented by the band, it is still possible that a mixed disorder could be present. For example, a patient with simultaneous development of severe metabolic alkalosis (say, from protracted vomiting) and severe metabolic acidosis (say, from concomitant diabetic ketoacidosis) could present with arterial blood gas values that are within normal limits.⁸ Similarly, if metabolic acidosis and metabolic alkalosis are simultaneously present, but one of the disorders is more severe, the blood gas results may lie squarely within the confidence band of the more severe disorder, concealing the presence of the opposing metabolic disorder. Although this is a limitation of the method, consideration of information from the history and screening laboratory testing, particularly the serum anion gap in this case, will usually overcome the diagnostic challenge. As another example, a patient with metabolic acidosis and simultaneous chronic respiratory alkalosis (Figure 28–1) may have blood gas values that are consistent with acute respiratory alkalosis. Again, the medical history and ancillary testing commonly surmount these limitations.

Each of the descriptive cardinal acid-base disturbances directly depicted in Figure 28–1 can be considered an acid-base diagnosis. Arriving at an acid-base diagnosis, however, is not an end in itself. Rather, the purpose of this identification process is to characterize the disorder so that a differential diagnosis can be formulated and the specific etiology then discerned. Pin-pointing the specific diagnosis requires additional information, available from the clinical context of the case, including the patient's medical, surgical, and medication use history, physical examination, and sometimes simple laboratory test results such as the serum potassium level, serum anion gap, or urine chloride concentration. Once the specific underlying diagnosis is ascertained, appropriate treatment can be implemented.