Background The traditional Henderson-Hasselbalch approach hasproven to be imprecise in critically ill patients. Stewart’s approachcan detect metabolic acidosis missed by traditional approach,including acidosis caused by increased unmeasured agreement(UA). The complexity of Stewart’s method leads to developmentof simpler modifications, simplified Fencl-Stewart and Figge-Stewart method. Agreement between both modifications isunknown.Objective This study aimed to measure the agreement of simplifiedFencl-Stewart with Figge-Stewart method in diagnosing metabolicacidosis in critically ill children.Methods The was performed in Hasan Sadikin General Hospital,Bandung from July to August 2006, involving <14 year-old criticallyill children. Blood samples for gas analysis, sodium, potassium,chloride and albumin measurement were taken simultaneously. Testresult was analyzed with simplified Fencl-Stewart and Figge-Stewartmethod and recorded with Excell spreadsheet. PASS was used forinterim analysis and DAG_Stat for raw agreement indices andKappa calculations.Results Forty-five (31 males, 14 females) children were enrolled.Acid base disturbances based on Stewart’s method were identifiedin 10 subjects with normal base excess and nine with normalbicarbonate. Significant increase of UA was detected in 11 of 45subjects with simplified Fencl-Stewart method, compared to thatof 12 subjects with Figge-Stewart method. Raw agreement indicesshowed 95.65% and 98.51% agreement for positive and negativeresult, Kappa was 0.94 (P=0.0000).Conclusions Excellent agreement is shown between simplifiedFencl-Stewart and Figge-Stewart method in diagnosing metabolicacidosis in critically ill children. Increased UA can be assessedwith both methods.