Anion gap normal range: low, normal and high values

What is the normal anion gap range?

The normal anion gap range depends on how the laboratory measures sodium, chloride and bicarbonate, so no single figure fits every report. On the ion-selective electrode analyzers that most laboratories now use, the range quoted most often is 3 to 11 mmol/L without potassium. Winter and colleagues found that range on a Beckman ASTRA analyzer in 1990, and in 2013 Sadjadi's group measured a mean of 7.2 ± 2 mmol/L, range 3 to 11, in 409 adults with normal kidney function and albumin. The older textbook value, 12 ± 4 mmol/L, or 8 to 16, comes from flame photometry in the 1970s. Sodium, chloride and bicarbonate are all monovalent ions, so mmol/L and mEq/L give the same number and a range printed in either unit reads the same way. The range to use for a patient is the one printed on that laboratory's report, and the anion gap calculator lets you replace its 3 to 11 default with it.

Why the anion gap normal range differs between labs

The anion gap normal range differs between laboratories mainly because analyzers measure chloride differently. Ion-selective electrodes read chloride higher than flame photometry and colorimetric methods did, and a higher chloride gives a smaller gap. Sadjadi and colleagues cite Winter's data: hospital employees tested on the ASTRA analyzer had an average chloride of 107 mmol/L, against a reference chloride of 101 mmol/L on the older Technicon instrument. Lolekha's group added that continuous-flow analyzers lose more carbon dioxide from the sample, which lowers the measured bicarbonate and widens the gap. Modern analyzers still disagree with each other: Roberts and Johnson found ranges of 5 to 10 mmol/L on a Synchron CX3 and 9 to 14 on a Hitachi 717. People differ as well. Kraut and Nagami note that the normal range in a population can span 8 to 10 mmol/L, so a patient's own earlier gap, measured when well, is the best baseline when one exists.

Sadjadi's group also warn that many laboratories never set their own anion gap range and use the manufacturer's suggested value, which they say could vary from 10 to 20 mmol/L. That is why the same gap of 13 can be flagged high on one report and normal on another.

Anion gap normal range with potassium

The anion gap with potassium is the older form of the formula, (Na⁺ + K⁺) − (Cl⁻ + HCO₃⁻). Sadjadi and colleagues explain that the gap was first calculated this way and later simplified to sodium minus chloride and bicarbonate, mainly because potassium varies over such a narrow range. Adding potassium raises the gap by exactly the potassium value, so a with-potassium range is the without-potassium range moved up by a typical potassium. Sadjadi's laboratory, for example, used a potassium reference range of 3.6 to 5.1 mmol/L, and its healthy adults averaged 4.3 mmol/L. The anion gap calculator follows the same logic: when you type a potassium, it moves its default range from 3 to 11 to 7 to 15 mmol/L, adding a typical potassium of 4 mmol/L. A gap calculated with potassium should never be compared with a range set without it, or a normal result will look about 4 mmol/L too high. Kraut and Nagami note that potassium matters most when it is very high or very low.

With sodium 140, potassium 4.0, chloride 106 and bicarbonate 24 mmol/L, the gap is 140 − (106 + 24) = 10 mmol/L without potassium and 144 − 130 = 14 mmol/L with it. Both are normal, each against its own range. Kraut and Nagami's example shows the other side: a potassium rising from 3.5 to 6.5 mmol/L adds an apparent 3 mmol/L to a gap that includes it.

Low, normal and high anion gap values

An anion gap result falls into one of three bands, and the edges of each band come from the laboratory's reference range rather than from fixed numbers. The anion gap calculator labels a value low when it is below the lower limit of the range entered, normal when it lies between the two limits, counting both ends, and high when it is above the upper limit. With its default ion-selective electrode range of 3 to 11 mmol/L without potassium, a gap below 3 is low, 3 to 11 is normal and above 11 is high; with potassium included, the default normal band becomes 7 to 15. Sadjadi and colleagues suggested that a gap below 2 mmol/L be treated as a low anion gap and a clue to paraprotein disorders or drug intoxication. The table below sets the bands side by side with what each one most often means, as Kraut and Madias summarise it.

What a high anion gap points to

A high anion gap most often means a metabolic acidosis in which an unmeasured acid anion has taken the place of bicarbonate. Mehta, Emmett and Emmett grouped the common causes in the GOLD MARK mnemonic, published in The Lancet in 2008 as an update to older lists: glycols (ethylene and propylene glycol), oxoproline (pyroglutamic acid, linked with acetaminophen), L-lactate, D-lactate, methanol, aspirin (salicylate), renal failure and ketoacidosis. Kraut and Nagami name the same core group in their 2013 review: lactic acidosis, ketoacidosis, toxic alcohols, salicylate intoxication and uremic acidosis. A high gap is not always an acidosis, though. Kraut and Madias list laboratory error, metabolic alkalosis, a high phosphate and a paraprotein as other causes, so the gap is read with the bicarbonate, the blood gas and the history. Once a high anion gap acidosis is established, the delta gap and delta ratio compare the rise in the gap with the fall in bicarbonate to look for a second disorder alongside it.

A normal gap does not rule these causes out. Kraut and Nagami report that only 58% of patients with a lactate above 2.5 mmol/L had an anion gap outside the normal range, and in Sadjadi's 299 patients with a lactate of 4 mmol/L or more, whose albumin averaged 2.4 g/dL, 27% had an uncorrected gap below 10.

What a low anion gap points to

A low anion gap most often has an ordinary explanation. Kraut and Madias list laboratory error and a low albumin as the commonest causes, because albumin carries most of the unmeasured negative charge in plasma and any fall in albumin shrinks the gap. Less often, a low gap is the first clue to something serious. A paraprotein can lower the gap, the classic example being the IgG of multiple myeloma, which adds unmeasured positive charge. Lithium is an unmeasured cation, and intoxication with it lowers the gap. Bromide and iodide are read as chloride by some analyzers, which inflates the chloride and can push the gap low or even below zero. In their 1977 review, Emmett and Narins also listed dilution, hypercalcemia, hypermagnesemia, hypernatremia and hyperviscosity among the causes of a reduced gap. Because laboratory error is the commonest cause and any error in the three or four measured values passes straight into the gap, repeating the electrolytes comes first for an unexplained low result.

Correcting the anion gap range for a low albumin

A low albumin lowers the anion gap, so the normal range only applies once the gap has been corrected for it. Figge and colleagues studied 152 critically ill patients and found that the observed gap underestimates the unmeasured anions by about 2.5 mmol/L for each 1 g/dL fall in albumin. The anion gap calculator adds that amount back: corrected anion gap = anion gap + 2.5 × (4.0 − albumin in g/dL), with albumin in g/L divided by 10 first. Kraut and Nagami give the effect as 2.3 to 2.5 mmol/L per g/dL. Some references take normal albumin as 4.4 or 4.5 g/dL instead of 4.0, which gives a corrected gap 1 to 1.25 mmol/L higher. The calculator's default example shows why it matters: sodium 140, chloride 106 and bicarbonate 24 mmol/L give a gap of 10, inside the 3 to 11 range, but with an albumin of 2.5 g/dL the corrected gap is 13.8 mmol/L, above it.

Compare the corrected value, not the raw one, with the laboratory range. Enter the albumin in the albumin-corrected anion gap calculator and it does both steps and labels the band.