The anion gap, the difference between routinely measured anions and cations of the serum, or NA⁺ – (Cl^– + HCO₃^–), is a useful tool for deciphering simple and mixed acid-base disorders. When low, the anion gap can be used to indicate the presence of occult cancer or drug poisoning. The traditional normal value for the anion gap has been 12 ± 4 mEq/L (normal range, 8 to 16 mEq/L), values that textbooks continue to reiterate [1, 2]. However, as noted by Winter and colleagues [3] and others [4], the range for the anion gap is much narrower than previously thought because of the use of the ion-selective electrode. Thus, we retrospectively analyzed the serum anion gap in 222 patients with normal serum albumin and creatinine levels. We attempted to eliminate the effects of renal failure and hypoalbuminemia, two factors known to modify anion gap values. The results (mean ±SD) of our analysis were as follows: serum sodium level, 140 ± 2.7 mEq/L; potassium level, 4.6 ± 0.4 mEq/L; chloride level, 105 ± 4 mEq/L; CO₂ level, 28 ± 2.9 mEq/L; and anion gap, 6.6 ± 2 mEq/L (range, 2.6 to 10.6 mEq/L). This range is almost identical to the value reported by Winter and colleagues [3].

The exact reason for the lower anion gap value is not entirely clear. One possible explanation may be that, with ion-selective electrode technology, serum chloride levels tend to be higher, whereas the serum sodium levels closely resemble serum water levels. These methods are widely used by clinical laboratories, some of which correct this chloride overestimation and use the typical anion gap range as a reference value. Knowledge of the lower range of the anion gap is not only of theoretical interest; it also has practical bedside applications. In a recent study of lactic acidosis in critically ill patients [5], all patients with blood lactate levels higher than 10 mmol/L had anion gap values exceeding 16 mEq/L; 50% of patients with lactate levels between 5 and 9.9 mmol/L and 79% of those with lactate levels between 2.5 and 5 mmol/L had anion gap values less than 16 mEq/L. I believe that this finding results from the fact that the current anion gap value is excessive; if a lower range is adopted, nearly all cases of metabolic acidosis could be easily identified. Also of note, the use of ion-selective electrodes has made pseudohyponatremia obsolete, although use of the term continues.