COMPARISON OF TWO DIFFERENT MODALITIES OF BICARBONATE FOR DETERMINATION OF ANION GAP IN CRITICALLY ILL PATIENTS

OBJECTIVE: To compare anion gap estimated through measured and calculated bicarbonate modalities to be used interchangeably in critically ill metabolic acidosis patients. METHODS: This cross-sectional study was conducted at Rehman Medical Institute, Peshawar, Pakistan from September 2019 to March 2020.  Out of 390 critically ill patients, 200 cases of metabolic acidosis were selected by non-probability consecutive sampling technique. Measured and calculated bicarbonate values were obtained through Cobas-c 501© (Roche) using enzymatic method and Cobas-b 221© (Roche) blood gas analyzer respectively. Statistical analysis was done by using SPSS-23. RESULTS: Normal anion gap metabolic acidosis (NAG-MA) and high anion gap metabolic acidosis (HAG-MA) based on calculated bicarbonate levels was observed in 57 (28.5%) and 143 (71.5%) cases as compared to 55 (27.5%) and 145 (72.5%) cases based on measured bicarbonate levels respectively (p>0.45). A significant correlation (r=0.888 and 0.656, r2=0.788 and 0.431) (p<.001) was found between mean values of NAG-MA and HAG-MA respectively, when each was calculated through both modalities of bicarbonate. On applying Bland Altman plot, bias was 1.45±2.89 and -2.14±3.87mmol/L, Upper limit of agreement (LOA) was 7.13 and 5.46 for NAG-MA and HAG-MA, while lower LOA was -4.23 and -9.74 for NAG-MA and HAG-MA respectively. According to the model Bland Altman plot and Story & Postuie criteria, bias and the levels of agreement were not appropriate to conclude that both entities of anion gap could be used interchangeably.  CONCLUSION: Normal and high anion gap metabolic acidosis estimated by measured and calculated bicarbonate cannot be used interchangeably in critically ill patients.