Overview
The anion gap is a calculated estimate of the difference between routinely measured cations and anions in plasma. It provides an indirect window on ions that are not included in a standard electrolyte panel, particularly negatively charged proteins, phosphate, sulfate, organic acids and other unmeasured substances. Clinicians most often use the value while assessing metabolic acidosis, but it can also reveal analytical problems, changes in plasma proteins or the presence of unusual cations and anions. The calculation is simple; the interpretation is not. A meaningful assessment requires the local laboratory reference interval, knowledge of whether potassium is included, the albumin concentration, the patient’s acid-base state and the clinical context.
Historical development
The principle arose from electroneutrality: plasma contains equal total positive and negative charge even though routine chemistry measures only a selection of ions. As automated electrolyte measurement became widespread in the twentieth century, the difference between measured sodium and the sum of chloride and bicarbonate became a practical bedside index. The term “anion gap” is therefore a convention rather than a literal gap in electrical charge. Over time, clinicians recognised that assay technology, especially the move from flame photometry to ion-selective electrodes, altered normal ranges. This history explains why a remembered universal normal value is unsafe and why each laboratory should establish its own interval.
How the anion gap is calculated
The most common equation is sodium minus the sum of chloride and bicarbonate. Some laboratories include potassium, producing a numerically higher result. Bicarbonate may be reported as total carbon dioxide on a chemistry panel; in most clinical settings it is used as the bicarbonate estimate. Values should be drawn from the same sample and checked for unit consistency. The calculated result should be compared with the laboratory’s own reference interval, not a generic threshold. A change from the patient’s baseline can be clinically relevant even when the result remains inside the printed range.
Physiological basis
Albumin is the largest contributor to the normal unmeasured anion pool. Phosphate, sulfate and organic anions make smaller contributions. Unmeasured cations include calcium, magnesium and positively charged proteins. A rise in unmeasured organic acids consumes bicarbonate and leaves their conjugate anions in plasma, increasing the gap. By contrast, bicarbonate loss through the gastrointestinal tract or kidney is commonly balanced by chloride retention, producing a hyperchloremic or normal-gap metabolic acidosis. These patterns are useful organising principles, but real patients often have more than one process.
Albumin correction
Hypoalbuminaemia lowers the expected anion gap and may conceal an accumulation of unmeasured acids. Albumin correction is therefore particularly important in critical illness, liver disease, nephrotic states, malnutrition and systemic inflammation. Several correction equations exist; a commonly used approximation adds about 2.5 mmol/L for each 10 g/L that albumin lies below 40 g/L. The exact relationship varies and correction should not be treated as perfectly precise. The key clinical lesson is that an apparently normal gap in a markedly hypoalbuminaemic patient may still represent high-gap acidosis.
Raised anion gap metabolic acidosis
Important causes include lactic acidosis, ketoacidosis, advanced kidney failure and toxic or drug-related organic acids. Lactic acidosis may arise from impaired tissue oxygen delivery, severe infection, seizures, adrenergic stimulation, liver dysfunction or medicines. Ketoacidosis may be diabetic, alcoholic, starvation-related or associated with sodium-glucose cotransporter-2 inhibitors, sometimes with only modest hyperglycaemia. Kidney failure permits retention of sulfate, phosphate and other acids. Toxicological causes include methanol, ethylene glycol, salicylates and other exposures; urgent specialist toxicology advice is appropriate when suspected.
Normal-gap metabolic acidosis
A normal-gap acidosis is usually associated with bicarbonate loss or impaired renal acid excretion accompanied by chloride retention. Gastrointestinal losses include diarrhoea, high-output fistulae and some urinary diversions. Renal causes include renal tubular acidosis, early kidney dysfunction, hypoaldosteronism and medicines such as carbonic anhydrase inhibitors. Large-volume chloride-rich fluid administration can contribute. The urine anion gap or urine ammonium assessment may help distinguish gastrointestinal bicarbonate loss from impaired renal acid excretion, but these tests have limitations and must be interpreted carefully.
Mixed acid-base disorders
A patient can have simultaneous high-gap acidosis, normal-gap acidosis, metabolic alkalosis or a respiratory disorder. The delta gap or delta ratio compares the rise in anion gap with the fall in bicarbonate and may suggest an additional metabolic process. It should be used as a prompt rather than a definitive diagnostic test because baseline bicarbonate and baseline gap vary, albumin correction may be incomplete and timing affects the pattern. Always examine pH, partial pressure of carbon dioxide, bicarbonate, respiratory compensation and the clinical trajectory together.
Low or negative anion gap
A low result is less common and should first trigger confirmation and review for analytical or pre-analytical error. Hypoalbuminaemia is the most frequent physiological explanation. Increased unmeasured cations may occur with lithium toxicity, severe hypermagnesaemia or marked hypercalcaemia, although the latter usually has limited effect. Positively charged paraproteins can reduce the gap, so an unexplained persistent low value may justify assessment for monoclonal gammopathy. Halide interference or extreme hyperlipidaemia and hyperproteinaemia can also distort measured electrolytes.
Practical clinical approach
First confirm that metabolic acidosis is present and review the blood gas or chemistry bicarbonate. Calculate the gap using the local convention and correct for albumin when appropriate. Examine lactate, ketones, glucose, renal function, medication history and potential toxic exposure. Review chloride and recent fluid therapy. Compare with previous values and repeat unexpected measurements. In a deteriorating patient, do not delay treatment of shock, sepsis, diabetic ketoacidosis or poisoning while refining the arithmetic. The calculation supports, rather than replaces, urgent clinical assessment.
Common pitfalls
Frequent errors include mixing arterial blood-gas electrolytes with laboratory chemistry values, using an outdated normal range, forgetting albumin, assuming a normal gap excludes lactate or toxic alcohols, and overlooking a second acid-base disorder. The gap may normalise as ketoanions are excreted while hyperchloremic acidosis persists during recovery. A high lactate can occur with a modest gap, and a raised gap can persist for reasons unrelated to the current emergency. Interpretation should therefore be serial and hypothesis-driven.
Worked clinical reasoning
Consider a septic patient with bicarbonate 16 mmol/L, an anion gap at the upper end of the local range and albumin 20 g/L. Without correction, the result may appear only mildly abnormal. Albumin adjustment can reveal a substantial excess of unmeasured anions, supporting urgent measurement and trending of lactate and consideration of additional acids. In another patient with prolonged diarrhoea, bicarbonate may be similarly low but chloride rises proportionately and the corrected gap remains normal, directing attention toward gastrointestinal bicarbonate loss and volume depletion.
Monitoring and escalation
Trend the gap together with pH, bicarbonate, lactate, ketones, renal function and the patient’s perfusion. A falling gap may indicate clearance of organic acids, but improvement should be confirmed clinically because chloride accumulation can keep bicarbonate low. Escalate promptly for severe acidemia, haemodynamic instability, altered consciousness, suspected toxic ingestion, refractory ketoacidosis or rapidly worsening kidney function. Renal replacement therapy, antidotal treatment or critical care support may be required depending on the cause.
Frequently asked questions
Is an anion gap of 12 always normal?
No. Normal ranges vary with laboratory methods, potassium inclusion and albumin. Use the reported local interval and clinical context.
Can the gap be normal in lactic acidosis?
Yes. Early disease, hypoalbuminaemia, chloride changes and analytical variation can blunt the rise.
Should every gap be corrected for albumin?
Correction is most useful when albumin is reduced or when clinical suspicion is greater than the uncorrected result suggests.
Does a falling gap prove recovery?
No. It is one component of serial assessment and may be influenced by fluid therapy and renal excretion.
Key points for practice
- Use the laboratory-specific equation and reference interval.
- Correct for significant hypoalbuminaemia.
- Interpret the gap with pH, bicarbonate, lactate, ketones and renal function.
- Look actively for mixed acid-base disorders.
- Confirm unexpected low or negative results.
- Never allow calculation to delay urgent treatment.
Clinical pearls and professional practice
Use the result to answer a clearly defined clinical question. Calculations and thresholds are most useful when they are embedded in a structured assessment that includes history, examination, baseline risk, time course and response to treatment. A result should change a decision, refine a differential diagnosis or prompt monitoring; otherwise, repeating it without a plan may add noise rather than value.
When handing over care, communicate the actual value, the relevant reference interval or threshold, the trend and the interpretation. Avoid shorthand that can be misunderstood by another team. Where local protocols differ from a general reference, the local validated pathway takes precedence. Document uncertainty explicitly and identify what subsequent finding would confirm or refute the working diagnosis.
Equity and population context also matter. Diagnostic equations and thresholds may perform differently at age extremes, in pregnancy, in people with unusual body composition and in populations under-represented in validation studies. Clinicians should recognise these limitations without withholding appropriate investigation. Specialist or laboratory advice is valuable when a result does not fit the clinical picture.
Quality and safety checklist
- Confirm patient identity, sample timing and measurement method.
- Check units and the local reference interval.
- Compare with previous values and the clinical trajectory.
- Review medicines, comorbidities and reversible factors.
- Look for conditions in which the test or equation is unreliable.
- Record the interpretation, action and follow-up plan.
- Escalate immediately when the patient is unstable, regardless of a reassuring calculation.
Clinical interpretation should remain proportionate to the decision being made. Reassess when new information becomes available, and avoid allowing a single numerical result to outweigh clear evidence of deterioration or an alternative diagnosis. Multidisciplinary discussion improves safety when treatment carries substantial benefit and risk.
Worked interpretation framework
Begin by confirming that sodium, chloride and bicarbonate were measured on the same specimen and that the local laboratory equation is known. Establish whether the patient is acidotic using blood gas and chemistry data rather than assuming that a raised gap alone proves acidaemia. Compare the calculated gap with the local reference interval, then consider albumin correction. Finally, integrate lactate, ketones, renal function, osmolar gap, toxicological history and the respiratory response. This sequence prevents the common error of naming a biochemical pattern without identifying the underlying process.
When bicarbonate is reduced, comparing the rise in anion gap with the fall in bicarbonate can suggest a mixed metabolic disorder. This comparison is approximate because baseline values vary, albumin alters the expected gap and timing affects both components. A disproportionately large fall in bicarbonate may suggest an additional normal-gap acidosis, whereas a relatively small fall may indicate concurrent metabolic alkalosis. The calculation should prompt a focused review rather than act as a stand-alone diagnosis.
Sampling, analytical and pre-analytical considerations
Unexpected results should trigger a check for sample contamination, delay in processing, unusual assay interference and changes in measurement platform. Bicarbonate may fall in a sample exposed to air, while chloride or sodium errors can produce implausible gaps. Markedly negative gaps are uncommon and may reflect hypoalbuminaemia, laboratory error, paraproteinaemia or an excess of unmeasured cations. Repeating the panel and discussing the result with the laboratory is appropriate when the pattern is inconsistent with the patient.
Special populations and clinical contexts
Critical illness commonly combines lactic acidosis, renal impairment, chloride loading and hypoalbuminaemia, so a single label may be misleading. Pregnancy, paediatric practice and severe malnutrition also require age- or context-appropriate reference values. In diabetic ketoacidosis, the gap may close before all physiological abnormalities resolve, and hyperchloremic acidosis can emerge during treatment. In toxic alcohol exposure, the osmolar gap and anion gap evolve over time; neither excludes poisoning when used alone.
Escalation and management implications
Treatment is directed at the cause, not the calculated number. Urgent escalation is required when acidosis is severe, the patient is unstable, lactate is rising, toxic ingestion is suspected, ketonaemia is substantial or renal failure is progressing. Serial measurement can help assess response, but the frequency should match the clinical trajectory. The gap should never delay resuscitation, antidotal therapy, renal replacement consultation or source control when these are otherwise indicated.
Documentation and handover
Record the equation used, whether albumin correction was applied, the relevant values and the proposed cause. During handover, state whether the process is high-gap, normal-gap or mixed, whether it is improving and which confirmatory tests are outstanding. This is more useful than reporting an isolated number and supports safe continuity of care.