Renal disease represents an interesting example of the potential overlap between normal and elevated AG acidosis. The high AG in patients with advanced chronic kidney disease is usually a late finding and reflects a severe reduction in glomerular filtration rate (GFR). As the GFR falls below 20 to 30 mL per minute (plasma creatinine > 3 to 4 mg per dL), anions, such as sulfate and phosphate, that would normally be excreted by filtration are retained. With lesser degrees of renal dysfunction, however, metabolic acidosis appears primarily because H⁺ (HCl) secretion is reduced, with little or no effect on the AG.

The metabolically generated daily acid load on a typical American diet approximates 50 to 100 mEq. This acid, mainly sulfuric, is immediately buffered by NaHCO₃:

The excess sulfate is excreted in the urine. If glomerular and tubular function decline in parallel, then the H⁺ and the SO₄^2- are retained, producing metabolic acidosis with a high AG. If, however, there is more significant tubular dysfunction, the excretion of acid is diminished, but excretion of sulfate may be maintained due to reduced reabsorption. In the latter setting, the AG may not rise as the serum HCO₃^– concentration decreases.

Therefore, the decrease in plasma HCO₃^– (severity of acidemia) need not correlate with extent of the rise in AG in renal dysfunction. Typically, the plasma bicarbonate concentration is greater than 12 mEq per L in patients with uncomplicated CKD. A search for a second acid–base disorder is indicated when a lower HCO₃^– concentration is identified.

Lactic acidosis is probably the most common cause of severe metabolic acidosis encountered in the intensive care unit. The AG is always increased above baseline (normal lactate level
is < 1.0 mmol per L)³ because lactate does not appear in the urine until a higher plasma concentration (at least 6 to 8 mmol per L) is achieved. Lactate levels greater than 5 mmol per L are considered diagnostic of lactic acidosis, although levels between 2 and 5 mmol per L may be significant in the appropriate clinical circumstances [4]. Metformin, a biguanide commonly used in the treatment of type II diabetes mellitus, can cause lactic acidosis, particularly in patients who present with acute or chronic renal insufficiency. Hemodialysis has been used in the treatment of metformin-induced acidosis [5].

Most cases of lactic acidosis involve the L-isomer. By comparison, D-lactic acidosis, a disorder observed most commonly in patients with abnormal bowel anatomy, results in a rise in the AG, but the lactate level is normal [6]. D-Lactate is not detected by the usual lactate assay, which measures only L-lactate and a specific assay must be requested to diagnose this disorder.

Ketoacidosis occurs when acetoacetic acid and β-hydroxybutyric acid are overproduced by the liver (see Chapter 101 for a complete discussion). Acetone, a breakdown product of acetoacetic acid, is not an acid; as such, it does not contribute to the acidemia or to the increased AG observed in this disorder.⁴

Although ketoacidosis is generally associated with an elevated AG, loss of ketoanions in the urine, particularly during intravenous fluid therapy, may attenuate the expansion of the
AG. Once formed, ketones may be excreted in the urine before, under the influence of insulin, they can be metabolized back to HCO₃^–. Because the initially produced ketoacids titrate the plasma HCO₃^– concentration downward, the loss of urinary ketoanions (as sodium or potassium salts) is tantamount to the renal loss of HCO₃^–. The net effect is that a high AG metabolic acidosis is present before therapy in most individuals with ketoacidosis but may convert to a normal AG metabolic acidosis once saline repletion occurs and ketogenesis ceases [7].

Massive muscle breakdown is an important cause of metabolic acidosis with an increased AG. Acute kidney injury due to myoglobinuria can cause retention of anions (e.g., phosphate) that have been released from damaged myocytes.

The most common acid–base abnormality observed with salicylate intoxication is a respiratory alkalosis caused by direct stimulation of the medullary respiratory center. A pure metabolic acidosis owing to aspirin toxicity is uncommon. With moderate-to-severe salicylate intoxication, the AG increases as salicylic acid, not simply due to accumulation of salicylate in the blood, promotes formation of lactic acid. The consequence is a mixed respiratory alkalosis with a high AG metabolic acidosis.

Methanol and ethylene glycol ingestions require early diagnosis because prompt treatment may be lifesaving. Inhibitors of alcohol dehydrogenase such as ethanol and fomepizole are used for this latter purpose, with fomepizole being the preferred agent, if available. Either agent can limit the conversion of the alcohols to their more toxic metabolic products. Methanol or ethylene glycol ingestion as a cause for high AG metabolic acidosis is suggested by the history and physical findings (see Chapter 119). The turnaround time for measurement of these toxins may delay treatment. The detection of an osmolal gap is a relatively quick way of supporting the suspected diagnosis.

The osmolal gap refers to the difference between the plasma osmolality (P_Osm) measured by the laboratory and that calculated using the following formula:

Normally, the measured P_Osm is higher than the calculated value by 10 mOsm per kg. A larger osmolal gap indicates the presence of osmotically active substances not normally present. The most frequent causes of an increased osmolal gap are ethyl alcohol, isopropyl alcohol, ketones, lactate, mannitol, ethylene glycol, and methanol. If ethanol, lactate, or ketones cannot be identified in a patient with an AG metabolic acidosis with an osmolal gap, the diagnosis of ethylene glycol or methanol intoxication should be strongly suspected [8]. In the intensive care unit setting, a high osmolal gap acidosis has also been associated with the use of continuous high-dose infusions of lorazepam for more than 48 hours. Propylene glycol, which is used as a solvent for intravenous medications including lorazepam, has been implicated as the cause of the hyperosmolar metabolic acidosis in this scenario [9]. It is important to understand that the presence of an osmolal gap that results from an ingested alcohol may only be detected when P_Osm is measured in the laboratory by freezing-point depression [10]. Also, after the alcohol is metabolized, the osmolal gap may disappear.

Toluene (present in glue and metabolized to hippuric acid) is a rare cause of metabolic acidosis. The AG rises early and then returns toward normal, as hippurate is excreted by the kidneys, a process that is similar to the renal handling of ketones (see previous discussion) [11].

Rarely, acetaminophen administration in therapeutic doses can lead to an elevated AG metabolic acidosis in metabolically stressed individuals, including pregnant women. In this setting, reduced glutathione stores permit the generation of pyroglutamic acid (5-oxoproline) [12].

The infusion of amino acid solutions during hyperalimentation is an abundant source of hydrochloric acid (HCl). The development of a metabolic acidosis is more common in patients with renal insufficiency.

Oral administration of cholestyramine chloride reportedly occasionally also causes acidemia. This resin, which is sometimes used in the management of hypercholesterolemia, is nonresorbable and can act as an anion-exchange resin, exchanging its Cl^– for endogenous HCO₃^– and producing a metabolic acidosis. Sevelamer chloride, a compound used as a phosphorous binder in chronic kidney disease, has been associated with lower bicarbonate levels than in those patients treated with calcium-based binders. The mechanism of the metabolic acidosis is believed to be similar to cholestyramine [13].

Loss of HCO₃^– from the gastrointestinal tract or kidneys can lead to a reduction in the plasma HCO₃^– level. Bowel contents are alkaline compared to blood because HCO₃^– is added by pancreatic and biliary secretions. HCO₃^– is later exchanged for Cl^– in the ileum and colon. The result is that most alkali secreted into the gut lumen is reclaimed by the colon. Gastrointestinal losses of HCO₃^– (or HCO₃^– precursors such as lactate and acetate) are most commonly observed in patients with diarrhea so severe that colonic transit time is too rapid for
alkali reabsorption. At times, the resulting HCO₃^– losses can approach 40 mEq per L of stool. Less frequently, metabolic acidosis from HCO₃^– depletion is a result of pancreatic fistulae, biliary drainage, or a ureterosigmoidostomy. In the last circumstance, the excretion of acid (as NH₄Cl) urine directly into the colon permits the exchange of HCl for HCO₃^– because the colon is permeable to H⁺ and Cl^–, unlike the urinary bladder [14]. This problem does not usually occur with an ileal bladder.

Pancreatic HCO₃^– losses are also observed in essentially all patients with a pancreatic allograft anastomosed directly to the urinary bladder. Bicarbonate secreted into the bladder cannot be reabsorbed.

Renal bicarbonate losses can cause or contribute to acidemia in type 2 (proximal) renal tubular acidosis (RTA; Table 71.2),⁵ during recovery from ketoacidosis (see previous discussion), and in patients who are posthypercapnia. In patients with proximal RTA (Table 71.3), the normal reabsorptive threshold for HCO₃^– is reduced. As a result, HCO₃^– can no longer be reabsorbed at a rate adequate to maintain the normal plasma level of approximately 25 mEq per L. As a consequence, the urine pH is alkaline (> 5.3), and the fractional excretion of HCO₃^– is elevated (> 15% of the filtered load).⁶ Normally, this value is less than 3% because more than 97% of HCO₃^– filtered through the glomerulus is reclaimed, primarily in the proximal tubule (Fig. 71.1). HCO₃^– wasting ceases, however, and the urine becomes acidic (pH < 5.3) once the plasma HCO₃^– concentration has stabilized at the new (lower) level. This process explains why the urine pH may be high or low in proximal RTA.

Renal HCO₃^– losses also occur as compensation for chronic respiratory alkalosis (chronic hypocapnia). During chronic hyperventilation, the blood pH increases as the PCO₂ decreases. As can be seen in Figure 71.1, an increase in intracellular pH diminishes H⁺ excretion, leading to a concomitant decrease in HCO₃^– reabsorption. These changes cause the plasma HCO₃^– concentration to fall, partially compensating for the alkalemia. If the stimulus for hyperventilation (e.g., hypoxemia) is suddenly eliminated, the PCO₂ rapidly returns to normal. Renal compensation, by comparison, continues for 1 to 2 more days, causing a persistent reduction in the plasma HCO₃^– concentration. The resulting posthypocapnic metabolic acidosis normally resolves spontaneously.