Skill in interpreting biochemical data to correctly identify and treat acid-base disorders is an essential competency for clinicians. Studies have, however, shown that proficiency in this area is lacking and that performance declines as disorders become more complex.1 This review presents a systematic approach and set of rules that should help clinicians to solve even the most challenging cases.

Systematic approach

Good history-taking and thorough clinical examination are both indispensable for providing clues to the nature and duration of the underlying disorder.

The first step in assessment of the arterial blood gas profile is appraisal of the pH: a reduced pH indicates acidaemia and an elevated pH indicates alkalaemia (Table 1). In simple acid-base disorders, compensation returns the pH toward, but seldom completely to, normal levels (except in mild chronic respiratory alkalosis).2 A normal pH does not always signify normal acid-base status, but may actually be a clue to the existence of a mixed disorder with components in opposing directions (Table 2).

The second step is appraisal of the pCO2 and [HCO3-] to identify the primary derangement and compensatory response. A low pH with elevated pCO2 and [HCO3-] is consistent with respiratory acidosis. A low pH with decreased pCO2 and [HCO3-] is consistent with metabolic acidosis. A high pH with decreased pCO2 and [HCO3-] is consistent with respiratory alkalosis. A high pH with elevated pCO2 and [HCO3-] is consistent with metabolic alkalosis. It is important to note that in simple acid-base disorders deviations in pCO2 and [HCO3-] are in the same direction. In mixed acid-base disorders deviations in pCO2 and [HCO3-] are in opposite directions.2

The third step entails assessing the adequacy of the compensatory response by applying the rules of compensation (Table 3). If insufficient time has elapsed to allow for complete compensation, a partially compensated simple acid-base disorder may be misclassified as a mixed one.2

The fourth step is to examine the serum electrolytes and anion gap (AG) and to decide whether additional testing is required, e.g. measurement of serum creatinine, plasma lactate or glucose and urinary ketones. Calculation of the delta ratio may assist with detection and characterisation of mixed acid-base disorders (see below). Urinary electrolytes and osmolal gap may be useful in assessment of metabolic alkalosis and normal AG metabolic acidosis.

The serum anion gap

In order to maintain electroneutrality, the sum of total circulating cations and anions must be equal. However, the formula for the AG routinely incorporates only the major ions: [Na+] – ([Cl-] + [HCO3-]) or ([Na+] + [K+]) – ([Cl-] + [HCO3-]). [HCO3-] is usually derived from total CO2 content in serum from venous blood, yielding values that are 2 - 3 mmol/l higher than in arterial blood.2 In reality, the AG reflects the difference between cations and anions unaccounted for (‘unmeasured’) by the formula. Since ‘unmeasured’ anions usually exceed ‘unmeasured’ cations, the reference interval for the AG is typically 5 - 12 mmol/l ([K+] excluded) or 9 - 16 mmo/l ([K+] included).3 The latter ranges, based on measurements by modern ion-selective electrodes, are lower than the 8 - 16 mmol/l ([K+] excluded) quoted for older flame photometry methods.4 In light of the wide range of normal values, a significant change in the AG of an individual patient may be easier to detect by comparing the value to a baseline obtained while no acid-base disorder was present.5

Decreased anion gap

Because circulating proteins are a major component of serum ‘unmeasured’ anions, hypoalbuminaemia decreases the AG by 2.5 mmol/l for every 10 g/l that albumin is below 40 g/l.2 Correction of the AG may be required to unmask a high AG metabolic acidosis, e.g. lactic acidosis, presenting with a normal AG due to hypoalbuminaemia. Other important causes of low or even negative AGs are laboratory error (measurements of [Na+] too low or [Cl-] and/or [HCO3-] too high) and overproduction of cationic proteins in IgG myeloma or polyclonal gammopathy.5

Increased anion gap

Elevated AGs are more common than decreased AGs and are generally caused by overproduction or decreased excretion of acid. Rarer causes of an elevated AG are metabolic alkalosis, severe hyperphosphataemia and overproduction of anionic paraproteins by IgA myeloma.5 Spurious increases may also be seen due to loss of water and CO2 from sera left exposed to air (usually ≤6 mmol/l after 2 hours).2 Metabolic alkalosis, particularly when it is chloride-responsive, may increase the AG by 4 - 6 mmol/l.2 Contributing factors include: raised albumin concentration (due to dehydration) and increased net anionic charge of proteins and enhanced lactate production by stimulation of glycolysis (the latter both the result of alkalaemia).5 , 6 Exogenous administration of phosphate resulting in serum concentrations of 6.0 - 7.5 mmol/l has been associated with AGs of >50 mmol/l.5

High anion gap metabolic acidosis

The greatest clinical utility of the AG, however, is in the differential diagnosis of metabolic acidosis. When an acid (AH) accumulates in the blood, the hydrogen ion [H+] is buffered by HCO3- and the retained acid anion [A-] contributes to unmeasured anions, raising the AG.6 The most common causes of high AG metabolic acidosis are: diabetic or alcoholic ketoacidosis, lactic acidosis, uraemic acidosis and intoxication with alcohol, methanol, ethylene glycol, salicylate, or carbon monoxide (Table 4). The offending anion can be identified with relative ease in most cases when the AG exceeds 30 mmol/l, declining to <70% of cases when the AG is <24 mmol/l.5 Of note is that the nitroprusside method used by dipsticks for detection of ketones reacts only with acetoacetate and not β-hydroxybutyrate. Since the latter often predominates in patients with alcoholic and diabetic ketoacidosis, the dipstick may significantly underestimate the degree of ketosis.2 In addition to screening tests for toxins, calculation of the osmolal gap (OG) as osmolality – (2[Na+] + [urea] + [glucose]), may be helpful as an indicator of intoxication. An OG >10 suggests the presence of an osmotically active substance (such as ethanol, methanol, or ethylene glycol) that is detected by measured but not calculated osmolality. Lesser elevations may be found in lactic acidosis and chronic renal failure.2 In cases of intoxication where blood collection is delayed, only mild increases in the OG may be observed due to conversion of alcohols to toxic metabolites.2

Normal anion gap metabolic acidosis

Normal AG metabolic acidosis is caused either by renal or gastrointestinal losses of bicarbonate or by addition of hydrochloric acid to the blood.5 The latter occurs when administered ammonium chloride or chloride salts of amino acids in hyperalimentation are metabolised to HCl by the liver.5 In the former conditions, loss of bicarbonate in the urine or stool along with sodium produces volume contraction, stimulating proximal renal tubular absorption of NaCl. Sodium bicarbonate losses are therefore replaced by sodium chloride, leading to a hyperchloraemic acidosis with an unchanged anion gap.5 However, if acidaemia is severe, the AG may actually drop by up to 4 mmol/l as a result of an increment in the net cationic charge of proteins.5 If hyponatraemia is present, the [Cl-] may be within the normal range. Normal AG metabolic acidoses can be classified into those in which serum [K+] is normal or increased and those in which it is low (Table 5).

Urine electrolytes

Urine chloride measurement is used to divide metabolic alkaloses into those that are chloride-responsive (U-[Cl-] <25 mmol/l) and those that are chloride-resistant (U-[Cl-] >40 mmol/l).1 In the former, the alkalosis is corrected by volume expansion using saline, whereas in the latter it is not. In metabolic alkalosis due to vomiting or nasogastric drainage urine chloride is used instead of sodium to assess volume status, because U-[Na+] may be high despite volume contraction. This is because the early phase of these conditions is associated with a significant bicarbonate diuresis that causes obligate urinary losses of sodium.4 Causes of metabolic alkalosis are presented in Table 6.

Measurement of urine electrolytes and osmolality with calculation of urine anion and osmolal gaps may be useful in the differential diagnosis of normal AG metabolic acidoses. Urine [Na+] and [K+] are expected to be low in diarrhoea, high in renal tubular acidoses and divergent (high [Na+] and low [K+]) in aldosterone deficiency or resistance.2 In acidosis of extrarenal origin the appropriate renal response is secretion by the collecting tubules of H+ ions, which combine with NH3 to be excreted as ammonium (NH4+).6 The urine AG, defined as [Na+] + [K+] – [Cl-], is an indirect way of estimating urinary NH4+ excretion. Since NH4+ is usually accompanied by Cl-, a negative urine AG (in the range of -20 to -50 mmol/l) indicates appropriate elevation of urinary NH4+ excretion in the face of acidosis.6 A positive urine AG (in the range of 20 - 30 mmol/l) indicates that impaired secretion of H+ ions, as found in distal renal tubular acidosis, plays a role in the evolution of the acidosis.2 Estimation of NH4+ excretion by the urine AG is unreliable when the urine contains significant amounts of bicarbonate or unusual anions, such as ketoacids or drugs, including penicillin or salicylate.6 In such cases, the urine osmolal gap, defined as urine osmolality – (2[Na+ + K+] + [urea] + [glucose]) should be used, since the gap mainly reflects the excretion of NH4+.2 In an acidotic patient, the urine OG should be between 150 and 200; values of 50 - 100 are indicative of impaired NH4+ excretion, consistent with distal renal tubular acidosis.2

Delta ratio

The ratio between the increase in AG (DAG) and the decrease in bicarbonate (DHCO3) is called the delta ratio. It is calculated as DAG/DHCO3- where:

DAG = patient’s AG (corrected for hypoalbuminaemia if present) – laboratory mean ‘normal’ value; and

DHCO3- = laboratory mean ‘normal’ value of 24 mmol/l – the patient’s total CO2 content.

The ratio is useful for detecting complex acid-base disorders that include a component of high AG metabolic acidosis. In high AG metabolic acidoses the premise is that total CO2 will decrease by 1 mmol/l for every 1 mmol/l increase in AG, i.e. a delta ratio of 1:1. However, in practice ratios between 1 and 2 may be observed, because hydrogen ions are not only buffered by extracellular bicarbonate.7 This is particularly true of lactic acidosis in which ratios are typically between 1.6 and 1.8.1 In diabetic ketoacidosis the ratio is closer to 1, because of loss of ketoacid anions in the urine.1 , 7 A delta ratio between 0.3 and 0.7 indicates that the increase in AG is small compared with the decrease in bicarbonate, suggesting co-existence of normal and high AG metabolic acidoses, e.g. diarrhoea and lactic acidosis.6 A delta ratio of >2 indicates that the increase in AG is large compared with the decrease in bicarbonate, suggesting co-existence of metabolic acidosis and metabolic alkalosis (see Case 2 below).7

Illustrative cases

Case 1

An 18-year-old woman with anorexia nervosa has results as depicted in Table 7. The elevated pH and [HCO3-] signify the presence of metabolic alkalosis. The pCO2 is appropriately elevated for the increase in bicarbonate [40 mmHg + (0.6 x (40-24)) = 40 mmHg + 9.6 = 49.6 mmHg], indicating adequate respiratory compensation. The urine electrolytes indicate volume contraction with secondary hyperaldosteronism. The corrected AG of 16 is high normal (13.4 + [(40-28)/10 x 2.5]), consistent with chloride-responsive metabolic alkalosis.

Case 2

A 55-year-old woman with type 1 diabetes mellitus is admitted to the Casualty Department in a semi-coma (Table 8). Her chronic medication includes insulin, digoxin and a thiazide diuretic. Although the arterial blood gas profile is unremarkable except for mildly decreased [HCO3-], the AG suggests the presence of a high AG metabolic acidosis, probably due to diabetic ketoacidosis. Calculation of the delta ratio [(31 – 12)/(24 – 22)] yields a value of 9.5, indicating mixed metabolic acidosis and alkalosis, the latter due to chronic diuretic therapy. The pCO2 is appropriately decreased for the bicarbonate [40 mmHg – (1.3 x (24-20)) = 40 mmHg – 5.2 = 34.8 mmHg], indicating adequate respiratory compensation.

References available at www.cmej.org.za

In a nutshell

• Recognition and appropriate treatment of acid-base disorders requires a systematic approach and application of a set of rules.

• Information obtained from the history and physical examination is essential for correct interpretation of biochemical data.

• The systematic approach comprises the following steps:

• appraise the pH: pH <7.35 indicates acidaemia; pH >7.45 indicates alkalaemia; normal or near-normal pH may indicate mixed acidosis and alkalosis

• identify the primary disorder and compensatory response by examining the pCO2 and HCO3-; deviations are in the same direction in simple disorders, but in opposite directions in mixed disorders

• assess whether compensation is adequate by applying the rules of compensation, keeping in mind the time interval required for compensation to be complete

• assess the serum electrolytes and anion gap; calculate the delta ratio if DAG and D HCO3- appear discordant

• define the acid-base disorder and request further testing if required, e.g. toxic screen, osmolal gap, lactate, ketones.

• Measurement of urine electrolytes and calculation of anion and osmolalgaps may be helpful in the differential diagnosis of metabolic alkaloses and normal anion gap metabolic acidoses.