High anion gap: what it means and when it matters

A high anion gap matters most when bicarbonate (often printed as CO2) on the same report is low. The gap is worked out from three numbers on your metabolic panel: sodium minus (chloride plus bicarbonate). Most of the gap is albumin and other charged molecules the panel does not count ¹. When the gap rises, something extra is taking up that space, usually an acid, and an acid uses up bicarbonate as it builds.

CalculatorAnion gap calculatorType in your sodium, chloride, bicarbonate and albumin to get your gap, corrected for albumin.Work it out

What counts as a high anion gap

Anion gap is reported in mEq/L (milliequivalents per litre, the unit for charged salts in blood). When labs switched to newer machines that measure each salt directly with a sensor, reported chloride ran higher and the normal gap fell from the old 8 to 16 mEq/L to roughly 3 to 11 ¹. Most reports now print about 3 to 11 or 4 to 12, some still print 8 to 16, and labs that add potassium into the formula use about 10 to 20. Use the range printed next to your result, and compare results from the same lab. The anion gap page explains the formula in full.

Low albumin pulls the gap down, so a normal-looking gap can hide an acid. Each 1 g/dL fall in albumin lowers the gap by about 2.5 mEq/L ¹, so doctors add 2.5 times (4.0 minus your albumin) to get a corrected gap. With an albumin of 2.0 g/dL, a reported gap of 10 corrects to 15, which is high. The anion gap calculator does this sum for you, and albumin has its own page.

High gap with low bicarbonate vs high gap with normal bicarbonate

Read a high anion gap beside bicarbonate. Bicarbonate is the blood's main acid buffer: when an acid builds up, bicarbonate is used up neutralising it, and the acid's leftover part widens the gap. So a high anion gap with low CO2 (bicarbonate) is the pattern that matters, because a real acid problem moves both numbers at once.

A high gap with a normal bicarbonate is the result most people with a routine panel see. Dehydration is the commonest reason: less water concentrates albumin, which nudges the gap up a little. If your bicarbonate is low too, the low bicarbonate page puts the mild end plainly: "A single 20 or 21 with everything else normal is common and often comes from a recent stomach bug, fasting, hard exercise or an underfilled tube." A lower value, or any low value with symptoms, is a different matter.

What causes a high anion gap (GOLDMARK, and why MUDPILES is out of date)

Medical students learned the causes as MUDPILES: methanol, uraemia (kidney failure), diabetic ketoacidosis, paraldehyde, iron or isoniazid, lactic acidosis, ethylene glycol and salicylates (aspirin). In 2008 three doctors proposed GOLDMARK instead, dropping causes that are now rare or obsolete (paraldehyde, isoniazid, iron) and adding propylene glycol, 5-oxoproline and D-lactate ². GOLDMARK stands for glycols, oxoproline, L-lactate, D-lactate, methanol, aspirin, renal failure and ketoacidosis. In practice four groups cover almost every case.

Lactic acidosis is the most common cause of metabolic acidosis in hospital, usually from sepsis or shock, and it is a marker of serious illness ⁵. Adult DKA is now defined by glucose at or above 200 mg/dL (or known diabetes), beta-hydroxybutyrate (the main ketone, measured in blood) at or above 3.0 mmol/L, and a blood pH (how acidic the blood is) below 7.3 or bicarbonate below 18 mmol/L ⁴. The same consensus warns that SGLT2 inhibitor drugs, such as empagliflozin and dapagliflozin, can cause ketoacidosis with a near-normal glucose, so a normal glucose does not rule DKA out ⁴. The diabetes page covers the condition itself.

Kidney disease raises the gap late. In earlier stages, kidney acidosis is usually the normal-gap kind; the gap climbs in advanced disease, when the kidneys can no longer clear phosphate, sulphate and urate. So a mild high gap in someone with an eGFR (your kidney filtration score) of 60 points somewhere other than the kidneys. If you only have creatinine, the eGFR calculator works it out, the chronic kidney disease page explains the stages, and Kidney function explained covers the whole renal panel. A high creatinine result has its own page.

Everyday worries: keto, fasting, exercise, alcohol and the blood tube

• A keto diet or a day of fasting raises ketones, but nutritional ketosis rarely lowers bicarbonate enough to matter in people without diabetes.
• If you take an SGLT2 inhibitor (such as empagliflozin or dapagliflozin), a keto diet or fasting can trigger ketoacidosis even with normal glucose. Talk to your prescriber before starting either, and get same-day care if you feel sick, vomit or breathe fast.
• Very hard exercise lifts lactate for an hour or so; a sample taken soon after a heavy session can show a slightly high gap that clears on a rest-day repeat.
• A binge with little food, followed by vomiting, can cause alcoholic ketoacidosis, which needs hospital care.
• A blood tube left uncapped or sitting too long before testing can lose CO2, which lowers bicarbonate and raises the calculated gap.

How doctors work out the cause of a high anion gap

1. 1Confirm there is an acidosisA low bicarbonate on the panel, sometimes with a blood gas to measure pH, separates a true acidosis from a lab or hydration effect.
2. 2Correct the gap for albuminThe albumin correction described above stops a low albumin from hiding a raised gap.
3. 3Check the four big causesGlucose and ketones for ketoacidosis, lactate for lactic acidosis, creatinine and eGFR for kidney failure.
4. 4Look for a toxin if nothing fitsDoctors compare measured and calculated serum osmolality (the osmolar gap, a count of dissolved particles). A raised osmolar gap points to methanol or ethylene glycol, though a normal one does not rule poisoning out.
5. 5Check for a second problemDoctors compare how far the gap rose with how far bicarbonate fell. If the two don't match, there may be a second acid or alkali problem as well ¹. They also check that breathing is offsetting the acid as expected (Winter's formula). Your doctor does these sums; you don't need to.

A worked example: sodium 138, chloride 102 and bicarbonate 20 give a gap of 138 minus 122, which is 16. With albumin of 3.0 g/dL, the corrected gap is 16 plus 2.5, which is 18.5. The bicarbonate of 20 is mildly low, so a doctor would read this as a mild high anion gap metabolic acidosis and look next at glucose, ketones, lactate and creatinine. Urine ketones and lactate each have their own page.

The gap also misses some acid: it can stay normal in a large share of people whose lactate is raised ¹. A normal gap in someone who is clearly unwell does not end the search; doctors measure lactate and ketones directly.

Symptoms of a high anion gap acidosis

A mild high gap found on a routine panel causes no symptoms at all. A real high-gap acidosis usually does: fast, deep breathing (the body blowing off carbon dioxide to offset the acid; doctors call it Kussmaul breathing), nausea and vomiting, belly pain, tiredness, and confusion or drowsiness as it worsens. In DKA, thirst and passing a lot of urine often come first. Most people with a significant acidosis feel ill enough to seek care before any blood test result comes back.

When to get help, and when a repeat test is enough

Beyond the emergency signs above, two situations call for a same-day or next-day call to your doctor even if you feel fine: a new high gap in someone with kidney disease, and a gap over about 20 with low bicarbonate.

For a gap a few points over range with a normal bicarbonate and no symptoms, a repeat at the same lab is the usual next step, at a time your clinician suggests, drinking normally beforehand and avoiding hard exercise the day before. Ask your clinician to include glucose and creatinine with eGFR on the repeat, and albumin if it was not on the first panel. Bring the earlier report so the two can be compared on the same range.

Bringing a high anion gap down: treat the cause

No supplement, diet or drink lowers the anion gap. The gap falls when the acid causing it is cleared: insulin and fluids in DKA, fluids, oxygen and infection treatment for lactic acidosis, food and fluids for alcoholic ketoacidosis, antidotes or dialysis for toxic alcohols. The lab-only kind from dehydration settles with normal hydration.

In chronic kidney disease, bicarbonate is the number doctors track, and treatment decisions are theirs to make. KDIGO's 2024 guideline puts the threshold this way ⁶:

In people with CKD, consider use of pharmacological treatment with or without dietary intervention to prevent development of acidosis with potential clinical implications (e.g., serum bicarbonate <18 mmol/l in adults).KDIGO 2024 Clinical Practice Guideline for CKD, Practice Point 3.10.1

Metformin-associated lactic acidosis is rare. Metformin is contraindicated below an eGFR of 30, and starting it is not recommended between 30 and 45 ³. A high gap in someone taking metformin is worth a call to the prescriber before changing it, unless you are vomiting or unwell, in which case get same-day care.

Medicines that bear on a high anion gap

Well-studied, rare riskMetforminA rare cause of lactic acidosis, mainly with poor kidney function; the page covers the eGFR limits and the labs to watch.

Electrolytes explained walks through sodium, chloride and CO2 together.

Free, no accountDrop in your lab report and see every marker explained in plain words, placed on the body, against your lab's range. It's read on your device: nothing is uploaded and nothing is kept.

Read my results Download the app