Serum anion gap in the differential diagnosis of metabolic acidosis in critically ill newborns

Abstract

Objectives: To determine in critically ill newborn infants (1) the range of the serum anion gap without metabolic acidosis and (2) whether the serum anion gap can be used to distinguish newborns with lactic acidosis from those with hyperchloremic metabolic acidosis. Study design: Umbilical arterial blood gases and serum electrolyte and lactate concentrations were measured simultaneously in 210 samples from 63 infants over the first week of life. Metabolic acidosis was defined as a blood base deficit (BD) >4 mmol/L. The anion gap was calculated as [Na +] – [C1 –] – [T CO 2]. Lactic acidosis was defined as a serum lactate concentration >2 SD above the mean serum lactate concentration in samples without metabolic acidosis. Results: In 89 blood samples with BD <4 mmol/L, serum lactate concentration decreased with postnatal age ( r = 0.51). The upper limit of serum lactate concentration was 3.8 mmol/L at less than 48 hours, 2.4 mmol/L between 48 and 96 hours, and 1.5 mmol/L for infants greater than 96 hours of age. The mean serum anion gap ± 2 SD in 174 samples without lactic acidosis was 8 ± 4 mmol/L; in 36 samples with lactic acidosis it was 16 ± 9 mmol/L ( P < .0001). Serum anion gap and lactate concentration were poorly correlated for samples without lactic acidosis ( r = 0.04) but highly correlated in those with lactic acidosis ( r = 0.81, P < .0001). None of the 85 samples with metabolic acidosis but without lactic acidosis had an anion gap >16 mmol/L; only 4 of 36 samples with lactic acidosis had an anion gap <8 meq/L. However, 25 of 36 samples with lactic acidosis had serum anion gaps of 8 to 16 mmol/L. Conclusion: In the presence of metabolic acidosis, a serum anion gap >16 mmol/L is highly predictive of lactic acidosis; a serum anion gap <8 is highly predictive of the absence of lactic acidosis; an anion gap = 8 - 16 mmol/L has no use in the differential diagnosis of metabolic acidosis in the critically ill newborn. (J Pediatr 1999;135:751-5)