Laboratory Assessment of Acid-Base Status

Evaluation of acid-base status requires analysis of both the arterial blood gas and an electrolyte panel. These collections should be obtained simultaneously or within a brief span. Use of a low-friction syringe allows ease of arterial puncture and collection. Excess heparin should be avoided to limit hemodilution and air bubbles removed from the syringe promptly to prevent gas exchange between blood and the air trapped in the syringe. Analysis should follow shortly. Mixed venous blood gas measurement is complicated by disassociation in arterial and venous PaCO₂, especially in the presence of poor tissue perfusion. On-line continuous monitoring of blood gas values may offer advantages in the future but is not yet generally available. The anion gap (AG) should be calculated from the electrolyte panel in every instance because it may reveal a high-AG metabolic acidosis, even in the setting of a mixed disorder where arterial pH, bicarbonate, or PaCO₂ may be in the normal range.

Normal Acid-Base Homeostasis

Systemic arterial pH is maintained between 7.35 and 7.45 by extracellular and intracellular chemical buffering together with respiratory and renal regulatory mechanisms. The control of PaCO₂ by the central nervous system (CNS) and respiratory systems and the control of the plasma bicarbonate by the kidneys stabilize the arterial pH by excretion or retention of acid or alkali. The metabolic and respiratory components that regulate systemic pH are described by the Henderson-Hasselbalch equation:

Under most circumstances, CO₂ production and excretion are matched, and the usual steady-state PaCO₂ is maintained at 40 mm Hg. Primary changes in PaCO₂ can cause acidosis or alkalosis, depending on whether PaCO₂ is above or below the normal value (respiratory acidosis or alkalosis, respectively). Underexcretion of CO₂ produces hypercapnia, and overexcretion causes hypocapnia, both of which will affect the systemic pH. The PaCO₂ is regulated primarily by neural respiratory factors and is not subject to regulation by the rate of CO₂ production; therefore, hypercapnia is usually the result of hypoventilation rather than of increased CO₂ production. Increases or decreases in PaCO₂ may represent primary derangements of the ventilatory function of the lungs (under neural respiratory control) or may be due to compensatory changes in response to a primary alteration in the plasma [HCO₃⁻].¹

A decrease in systemic pH is termed acidemia, whereas an increase in pH is called alkalemia. Conversely, such changes in pH can occur with changes in PaCO₂ or serum bicarbonate, which are referred to as alkalosis or acidosis. An example of a simple disorder would be a patient with acute pancreatitis with profound vomiting (loss of acid) who will have alkalemia (pH 7.49) due to metabolic alkalosis, with bicarbonate of 35 mEq/L and a respiratory acidosis (PaCO₂ 47.5 mm Hg) as compensatory response (see later discussion).

Primary alteration of PaCO₂ evokes two metabolic mechanisms to limit change in systemic pH: the fast-acting cellular buffering and the renal-adaptive response, a slower process that becomes more efficient with time. This metabolic response would be secondary (or compensatory) to the primary respiratory disorder. A primary change in the plasma [HCO₃⁻] as a result of metabolic or renal factors results in compensatory changes in ventilation that blunt the changes in blood pH that would occur otherwise. Such respiratory alterations are referred to as secondary or compensatory changes because they occur in response to primary metabolic alterations.¹

The kidneys regulate plasma [HCO₃⁻] through three main processes: (1) “reabsorption” of filtered HCO₃⁻, (2) generation of “new” HCO₃⁻, which is accomplished by formation of titratable acid, and (3) excretion of NH₄⁺ in the urine. The kidney filters approximately 4000 mEq of HCO₃⁻ per day, and between 80% and 90% of HCO₃⁻ is reabsorbed in the proximal tubule. The distal nephron reabsorbs the remaining HCO₃⁻ and more importantly, secretes protons generated from dietary protein intake to defend systemic pH. Metabolism of the average diet rich in protein produces fixed acids that consume bicarbonate on entry into the extracellular fluid. Although the quantity of protons from dietary protein metabolism is small (40-60 mEq/day), it must be secreted to prevent chronic positive H⁺ balance and metabolic acidosis. This quantity of secreted protons (net acid) is represented in the urine as titratable acid and NH₄⁺. Metabolic acidosis in the presence of normal renal function augments net acid excretion by markedly increasing NH₄⁺ production and excretion. It is important to note that this vital compensatory mechanism is impaired in chronic renal failure, hyperkalemia, and renal tubular acidosis.^1,2

In sum, these regulatory responses—chemical buffering, regulation of PaCO₂ by the respiratory system, and regulation of [HCO₃⁻] by the kidneys—act in concert to maintain a systemic arterial pH between 7.35 and 7.45.

Simple Acid-Base Disorders

Primary respiratory disturbances (primary changes in PaCO₂) invoke compensatory metabolic responses (secondary changes in [HCO₃⁻]), and primary metabolic disturbances elicit predictable compensatory respiratory responses by causing changes in PaCO₂.

The degree of primary alteration and secondary compensation in either or both of these two variables (acidosis or alkalosis) determines the systemic pH (acidemia or alkalemia). For example, metabolic acidosis due to an increase in endogenous acids (e.g., ketoacidosis) lowers extracellular fluid [HCO₃⁻] and decreases systemic pH. This stimulates the medullary chemoreceptors to increase ventilation and to return the ratio of [HCO₃⁻] to PaCO₂, and thus pH, toward normal, although not to normal. Table 109-1 contains the acid-base disturbances along with the appropriate compensatory response for simple disorders. The degree of respiratory compensation expected in a simple form of metabolic acidosis can be predicted from the relationship PaCO₂ = (1.5 × [HCO₃⁻]) + 8 ± 2; that is, the PaCO₂ is expected to decrease 1.25 mm Hg for each mEq/L per liter decrease in [HCO₃⁻]. Thus, a patient with metabolic acidosis and [HCO₃⁻] of 12 mEq/L would be expected to have a PaCO₂ between 24 and 28 mm Hg. Values for PaCO₂ below 24 or greater than 28 mm Hg define a mixed disturbance (metabolic acidosis plus respiratory alkalosis or metabolic acidosis plus respiratory acidosis, respectively). A readily available (though not as reliable) method of determining the nature and degree of compensatory response is the use of nomograms (Figure 109-1).^1,4 If the arterial acid-base value falls within one of the shaded bands in Figure 109-1, one may assume that a simple acid-base disorder is present, and a tentative diagnostic category can be assigned. Values that fall outside the shaded area suggest the presence of a mixed disorder. These nomograms, though helpful, are not substitutes for an appreciation of the limits of compensation as displayed in Table 109-1.⁴

Figure 109-1 Acid base normogram. Shaded areas represent 95% confidence limits of normal respiratory and metabolic compensations for primary disturbances. Points outside shaded areas represent a mixed disorder, assuming absence of laboratory error.

Mixed Acid-Base Disorders

Mixed acid-base disorders, defined as independently coexisting disorders, not merely compensatory responses, are more often seen in patients in ICUs and can lead to dangerous extremes of pH. A patient with diabetic ketoacidosis (DKA; high-AG metabolic acidosis) may develop an independent and superimposed respiratory problem leading to respiratory acidosis or alkalosis. Patients with underlying pulmonary disease may not respond to metabolic acidosis with an appropriate ventilatory response because of insufficient respiratory reserve. Such imposition of respiratory acidosis on metabolic acidosis can lead to severe acidemia and a poor outcome. When metabolic acidosis and metabolic alkalosis coexist in the same patient, the pH may be normal or near normal. When the pH is normal, an elevated AG (see later) denotes the presence of a metabolic acidosis. Patients who have ingested an overdose of drug combinations such as sedatives and salicylates may have mixed disturbances as a result of the acid-base response to the individual drugs (metabolic acidosis mixed with respiratory acidosis or respiratory alkalosis, respectively). Even more complex are triple acid-base disturbances. For example, patients with metabolic acidosis due to alcoholic ketoacidosis may develop metabolic alkalosis due to vomiting and superimposed respiratory alkalosis due to the hyperventilation of hepatic dysfunction or alcohol withdrawal.¹ In general, a normal arterial pH in face of abnormal bicarbonate level, PaCO₂, or AG is highly suggestive of a complex and mixed acid base disorder.

Pathway to Diagnosis of Acid-Base Disorders

A stepwise approach to the diagnosis of acid-base disorders follows and is summarized in Table 109-2. In the determination of arterial blood gases by the clinical laboratory, both pH and PaCO₂ are measured, and the [HCO₃⁻] is calculated from the Henderson-Hasselbalch equation. This calculated value should be compared with the measured [HCO₃⁻] (or total CO₂) on the electrolyte panel. These two values should agree within 2 mEq/L. If they do not, the values may not have been drawn simultaneously, a laboratory error may be present, or an error could have been made in calculating the [HCO₃⁻]. After verifying the blood acid-base values, one can then identify the precise acid-base disorder.¹

TABLE109-2 Steps in Acid-Base Diagnosis

Adapted in part from DuBose TD Jr. Acid-base disorders. In: Brenner BM, editor Brenner and Rector’s the kidney 8th ed. Philadelphia: Saunders; 2008, p. 513.

The most common causes of acid-base disorders should be kept in mind while probing the history for clues about the etiology. For example, established chronic renal failure is expected to cause a metabolic acidosis, and chronic vomiting frequently causes metabolic alkalosis. Patients with pneumonia, sepsis, or cardiac failure frequently have respiratory alkalosis, and patients with chronic obstructive pulmonary disease or a sedative drug overdose often display a respiratory acidosis. The drug history is important because loop or thiazide diuretics may cause metabolic alkalosis, and the carbonic anhydrase inhibitor, acetazolamide, can result in metabolic acidosis.

Blood for electrolytes and arterial blood gases should be drawn simultaneously before therapy, because an increase in [HCO₃⁻] occurs with metabolic alkalosis and respiratory acidosis. Conversely, a decrease in [HCO₃⁻] occurs in metabolic acidosis and respiratory alkalosis.^1,2

Metabolic acidosis often leads to hyperkalemia as a result of cellular shifts in which H⁺ is exchanged for K⁺ or Na⁺. For each decrease in blood pH of 0.10, the plasma [K⁺] should rise by 0.6 mEq/L. This relationship is not invariable, however. DKA, lactic acidosis, diarrhea, and renal tubular acidosis are regularly associated with potassium depletion because of urinary K⁺ wasting.¹

By definition, a high-AG acidosis has two identifying features: a low [HCO₃⁻] and an elevated AG. This means that the elevated AG will persist even if another disorder coincides to modify the [HCO₃⁻] independently. In such a situation, one will be faced with an apparent normal bicarbonate level (and perhaps a normal pH) despite acid accumulation.

Anion Gap

Calculation of the anion gap is a key step in evaluation of acid-base disorders (see Table 109-2). Because normally the total unmeasured anions exceed the total unmeasured cations as indicated by the electrolyte panel, there exists an AG of 9 ± 3 mEq/L in plasma. The concentration of potassium in the blood usually is relatively small compared with that of sodium, chloride, and bicarbonate, so many clinicians omit this variable when calculating the AG.⁵

Simply put, AG represents unmeasured anions in plasma and is calculated as shown in the following equation. The various contributors to plasma AG in normal physiologic state and in metabolic acidosis are depicted in Figure 109-2. Patients with underlying pulmonary disease may not respond to metabolic acidosis with an appropriate ventilatory response because of insufficient respiratory reserve. Such imposition of respiratory acidosis on metabolic acidosis can lead to severe acidemia and a poor outcome. When metabolic acidosis and metabolic alkalosis coexist in the same patient, the pH may be normal or near normal. When the pH is normal, an elevated AG (see later) denotes the presence of a metabolic acidosis.

Figure 109-2 Contributors to plasma anion gap in normal physiologic state and in metabolic acidosis.

(Data from Gamble JL. Chemical anatomy, physiology, and pathology of extracellular fluid. 6th ed. Cambridge: Harvard University Press; 1954; and from Stewart PA. How to understand acid-base. New York: Elsevier; 1981.)

The unmeasured anions include predominately anionic proteins such as albumin but also phosphate, sulfate, and organic anions. An increase in the AG is most often due to an increase in unmeasured anions and less commonly is caused by a decrease in unmeasured cations (calcium, magnesium, potassium) (Table 109-3).¹ When endogenously produced acid anions such as acetoacetate and lactate accumulate in extracellular fluid, the AG increases, causing a high-AG acidosis. In addition, the AG may increase with an increase in anionic albumin, either because of increased albumin concentration due to profound volume depletion or alkalosis, which alters albumin charge (increased negative charge).⁶

TABLE109-3 Anion Gap in the Diagnosis of Metabolic Acidosis Anion Gap = Na⁺ − (Cl⁻ + HCO₃⁻) = 9 + 3 mEq/L

* Albumin is the major unmeasured anion. A decline in serum albumin of 1.0 g/dL from the normal value of 4.5 g/dL decreases the anion gap by 2.3-2.5 mEq/L. Correction is very important to diagnose anion gap acidosis in setting of hypoalbuminemia.

Adapted from Emmett M, Narins RG. Clinical use of the anion gap. Medicine 1997;56:38-54; from Oh MS, Carroll HJ. The anion gap. N Engl J Med 1977;297:814-7; and from Kraut JA, Madisa NE. Serum anion gap: its uses and limitations in clinical medicine. Clin J Am Soc Nephrol 2007;2:162-74. Epub 2006 Dec 6.

A low serum AG is not an uncommon occurrence and most frequently is the result of severe hypoalbuminemia. Albumin is the major contributor to serum AG, and a decline of 1 g/dL in serum albumin from the normal value of 4.5 g/dL will cause a reduction of 2.3 to 2.5 mEq/L in AG. Given the pivotal role of AG in formulating differential and treatment plans in acid-base disorders, it is extremely important to correct for low albumin when calculating AG in setting of hypoalbuminemia (see Table 109-3). Other causes of a low AG include:

Besides hypoalbuminemia, polyclonal gammopathy and monoclonal gammopathy with excessive accumulation of cationic immunoglobulin (Ig)G are the most common clinical disorders associated with a low serum AG. Therefore, once laboratory error and hypoalbuminemia have been excluded, a search for accumulation of IgG should be initiated. In patients with disturbed mentation or unexplained clinical findings, the possibility of lithium ingestion, bromism, or iodide intoxication should be considered. When the serum AG is negative in the absence of laboratory error, an extremely uncommon situation, bromide intoxication and iodide intoxication, should be excluded.⁵

In the face of a normal serum albumin, a high AG is usually due to non–chloride-containing acids that contain inorganic (phosphate, sulfate), organic (ketoacids, lactate, uremic organic anions), exogenous (salicylate or ingested toxins with organic acid production), or unidentified anions. As mentioned earlier, a high-AG acidosis has two identifying features: a low [HCO₃⁻] and an elevated AG. The latter is present even if an additional acid-base disorder is superimposed to modify the [HCO₃⁻] independently. Metabolic acidosis of the high-AG variety, concomitant with either chronic respiratory acidosis or metabolic alkalosis, represents a situation for which [HCO₃⁻] may be normal or increased. Nevertheless, the AG is elevated, signaling the presence of the acidosis (see Table 109-3). In a typical simple AG metabolic acidosis, one would expect an equal but reciprocal change in serum bicarbonate and the AG, but this relationship does not hold when mixed disorders are present. Therefore, ΔAG versus ΔHCO₃⁻ is an important tool to search for a concealed acid-base disorder. For example, the combination of metabolic acidosis and metabolic alkalosis is expected to be present in a patient with advanced renal failure with several days’ history of vomiting. This mixed disorder would be most easily recognized when the AG is elevated but the HCO₃⁻ concentration and pH are near normal (ΔAG > ΔHCO₃⁻). In general, a ΔAG/ΔHCO₃⁻ value of 1 is typical of pure high-AG acidosis such as lactic or DKA acidosis.⁷ A ratio significantly greater than 1 suggests the presence of metabolic acidosis and metabolic alkalosis, whereas a ratio less than 1 suggests the presence of mixed gap and non-gap metabolic acidosis. However, studies have indicated variability in the ΔAG/ΔHCO₃⁻.⁵ This observation undercuts the ability to use this ratio alone to detect complex acid-base disorders, thus emphasizing the need to consider additional information to obtain the appropriate diagnosis.

To further illustrate the importance of anion gap as the only clue to the presence of an acid-base disorder in face of normal values in the ABG, consider the following case: A 49-year-old male with a history of heavy alcohol consumption and poor dietary intake is admitted for persistent vomiting of several days’ duration. He is found to have metabolic alkalosis, with a pH of 7.55, PaCO₂ of 48 mm Hg, [HCO₃⁻] of 40 mEq/L, [Na⁺] of 135 mEq/dL, [Cl⁻] of 80 mEq/dL, and [K⁺] of 2.8 mEq/dL. If such a patient were then to develop a superimposed alcoholic ketoacidosis with a [β-hydroxybutyrate] of 15 mM, arterial pH would fall to 7.40, [HCO₃⁻] to 25 mEq/L, and PaCO₂ to 40 mm Hg. Although these blood gas findings are normal, the AG is elevated at 30 mEq/L, indicating a mixed metabolic alkalosis and metabolic acidosis.

Effects of Acidosis

The effects of acidemia on the body are multiple (Table 109-4). The fall in blood pH is accompanied by a characteristic increase in ventilation, especially the tidal volume (Kussmaul respiration). Intrinsic cardiac contractility may be depressed, but inotropic function can be normal because of catecholamine release. Both peripheral arterial vasodilation and central venoconstriction can be present; the decrease in central and pulmonary vascular compliance predisposes to pulmonary edema with even minimal volume overload. CNS function is depressed, with headache, lethargy, stupor, and in some cases, even coma. Glucose intolerance may also occur.¹

TABLE109-4 Systemic Effects of Acidosis

Lactic Acidosis

Lactic acidosis is one the most common causes of high-AG acidosis in the ICU. An increase in plasma L-lactate is most commonly due to increased production of lactate in setting of an imbalance in oxygen supply and demand at the tissue level (type A). Thus type A lactic acidosis is thought to be caused by tissue hypoperfusion and/or severe hypoxemia, although in recent years this view is thought to be an oversimplification. Non-hypoxic conditions can also generate significant lactic acidosis (type B) in a variety of clinical settings such as malignancies, hepatic failure, or ingestion of drugs/toxins. The following are some of the causes of lactic acidosis in clinical setting (also see Table 109-6):

Among the most common causes of lactic acid acidosis in medical ICUs is unrecognized bowel ischemia or infarction in a patient with severe atherosclerosis or cardiac decomposition receiving vasopressors.¹⁰ Moreover, independent of the cause of hemodynamic instability, use of catecholamines, especially epinephrine, also results in lactic acidosis, presumably by stimulating cellular metabolism such as hepatic glycolysis.⁹

D-Lactic acid acidosis is due to formation of D-lactate by gut bacteria and may cause both an increased AG and hyperchloremia (see Table 109-6).^9,10 This condition is caused by overgrowth of intestinal flora and may be associated with jejunoileal bypass or intestinal obstruction.

Lactic acidosis is among the most frequent and critical of all AG acidoses observed in the acute care setting. A study of 50 ICUs revealed an incidence of elevated lactate levels in over 60% of patients.¹¹ Whether lactic acidosis represents a unique entity or is a consequence of a variety of other conditions common to the ICU has been debated. Lactate concentrations are mildly elevated in nonpathologic states (such as exercise), but the magnitude of elevation is generally small. For purposes of definition, a serum L-lactate level greater than 4 mEq/L (normal being 1 mEq/L) is thought to represent a clinically significant lactic acidosis and is the initiating point for resuscitative protocols in many critical care units. Nevertheless, some patients in the ICU maintain serum lactate levels between 2 to 4 mEq/L, and it is uncertain whether such patients progress to frank lactic acidosis, but studies are suggestive of higher mortality for patients with even intermediate rises in serum lactate.¹²

L-Lactic acid is the product of the anaerobic metabolism of pyruvate, which is derived from glucose by means of the Embden-Meyerhof pathway. Under aerobic conditions, pyruvate is oxidized to acetyl CoA. In the absence of oxygen, however, pyruvate is reduced instead to lactate. Lactate is converted back to pyruvate by both the liver and the kidney via the Cori cycle, assuming normal liver function. Hepatic dysfunction, therefore, predisposes to the development of lactic acidemia in the presence of tissue hypoperfusion.

Whereas the lactic acid acidoses have been classified by Huckabee and Cohen into two types—type A (hypoxic) and type B (non-hypoxic) as noted earlier¹³—it has been recognized that lactic acidosis is often the result of the simultaneous existence of both hypoxic and non-hypoxic factors, and in many cases the precise etiology is difficult to establish. Decreased arterial perfusion to peripheral tissues in shock despite adequate arterial oxygen content, for example, results in L-lactic acid accumulation. Severe acidemia decreases portal blood flow and hepatic clearance of lactic acid.¹ Moreover, in sepsis there is both a decrease in tissue perfusion and a decrease in oxygen utilization. Technically, therefore, the classification of lactic acidosis is primarily of conceptual interest.

Numerous drugs have been implicated in the occurrence of lactic acidosis (see Table 109-6). Of particular note is biguanide (metformin) and antiretroviral therapy, specifically the nucleoside reverse transcriptase inhibitors (NRTIs) used in treatment of HIV infection. The biguanide family of hypoglycemic agents, which includes metformin, have been associated with mild elevation in serum lactate (usually <2 mEq/L) in patients with otherwise normal renal and hepatic function. Cases of severe lactic acidosis associated with use of metformin are typically accompanied by presence of sepsis and/or profound renal failure. The mechanism is largely unknown,¹⁴ and the current recommendations are that the drug not be used in patients with congestive heart failure, liver disease, and significant renal insufficiency (creatinine >1.5 mg/dL in men, or >1.4 mg/dL in women).^9,15 Nucleoside analogs used in treatment of HIV infections inhibit mitochondrial polymerase and lead to lactic acid accumulation. Hyperlactemia is common with NRTI therapy, especially stavudine and zidovudine, but the serum lactate is mildly elevated and well compensated. Risk factors for NRTI therapy–associated lactic acidosis include a creatinine clearance less than 70 mL/min and a low CD4⁺ T-lymphocyte count.^9,16

Carbon monoxide poisoning produces lactic acidosis by reducing the oxygen-carrying capacity of the hemoglobin, resulting in tissue hypoxia. A newly recognized and not too uncommon cause of lactate accumulation is propylene glycol. This agent is used as carrier for a variety of IV medications used in ICU setting, most notably diazepam, lorazepam, nitroglycerin, and etomidate. Metabolism of propylene glycol by alcohol dehydrogenase in the liver results in lactate formation, which is then converted to pyruvate and shunted to glycolytic pathways. There have been numerous reports of high-AG acidosis and elevated serum osmolarity in patients receiving benzodiazepine infusions in the critical care setting.¹⁷

Critically ill patients with a significantly elevated AG or low serum bicarbonate should be suspected of having a lactic acidosis, particularly in the presence of hepatic insufficiency. A high index of suspicion must be maintained, however, because the AG is a relatively insensitive reflection of lactic acidosis. Iberti and coworkers reported a poor correlation between arterial pH, the AG, and serum lactate levels.¹⁸ Fifty percent of patients with serum lactate levels above 5 and less than 9.9 mmol/L displayed a normal AG.

Several investigators have sought to characterize the prognostic value of serum lactic acid levels. Studies have found an inverse correlation between mortality and L-lactate levels above 2.0 to 2.5 mmol/L.^19–21 Prognosis is related to lactate concentration, as well as the ability to metabolize a lactic acid load after a resuscitative effort. Although lactic acidemia is associated with adverse outcome during critical illness, it does not appear to be a direct causative agent, rather a marker of poor prognosis. Therapeutic interventions that target the primary pathophysiology rather than the lactic acidosis per se have been shown to have outcome benefits.²² Moreover, normalization of elevated lactic acid levels regardless of causative source is associated with better outcome in patients with sepsis, but this may be a surrogate for the acuity of the illness and index of organ dysfunction.²³ Falk and associates observed that the ability to lower serum lactate levels by 50% within 18 hours after resuscitation correlated with a significantly greater rate of survival.²⁴ Other studies have reinforced these findings, revealing both significantly lower lactate levels and increased ability to clear lactate in survivors as opposed to nonsurvivors.²⁵ To add to this argument, dichloroacetate, which indirectly decreases lactic acid level, has not been shown to improve survival, suggesting that elevated lactate level is an epiphenomenon with varying prognostic value in different clinical settings.²⁶