An automated analyzer provides clinically concordant results to manual back titration for quantitation of bicarbonate in pancreatic juice

OBJECTIVES: Secretin pancreatic function tests play an important role in the diagnosis of chronic pancreatitis. Back titration is the standard method for measurement of bicarbonate in pancreatic juice but is time consuming and manually performed. Use of an autoanalyzer for this purpose is not validated. METHODS: Bicarbonate concentrations in secretin-stimulated pancreatic juice specimens were quantitated by manual back titration, a clinical chemistry autoanalyzer (automated bicarbonate, Roche Cobas c501, Roche Diagnostics, Indianapolis, Ind), and a blood gas analyzer (calculated bicarbonate, GEM 3000, Instrumentation Laboratories, Bedford, Mass). Kappa statistic analysis, Bland-Altman analysis, and Lin concordance correlation coefficients were calculated. RESULTS: Ninety specimens from 31 subjects were included. Using a bicarbonate concentration of 80 mEq/L as a cutoff value, there was poor agreement between back titration and calculated bicarbonate (κ = 0.188); however, only 1 specimen showed discrepancy between back titration and automated bicarbonate (κ = 0.977). The limit of agreement between the back titration and automated bicarbonate was -9.0 to + 8.3 mEq/L. The Lin concordance correlation coefficient between the 2 methods was 0.931 (P < 0.001). CONCLUSIONS: There is strong concordance between manual back titration and chemistry autoanalyzer methods for measurement of bicarbonate concentrations in pancreatic juice. Autoanalyzers may replace back titration for this purpose. Blood gas analyzers are unsuitable.