Cyst Fluid Glucose is Rapidly Feasible and Accurate in Diagnosing Mucinous Pancreatic Cysts

OBJECTIVES: Better diagnostic tools are needed to differentiate pancreatic cyst subtypes. A previous metabolomic study showed cyst fluid glucose as a potential marker to differentiate mucinous from non-mucinous pancreatic cysts. This study seeks to validate these earlier findings using a standard laboratory glucose assay, a glucometer, and a glucose reagent strip. METHODS: Using an IRB-approved prospectively collected bio-repository, 65 pancreatic cyst fluid samples (42 mucinous and 23 non-mucinous) with histological correlation were analyzed. RESULTS: Median laboratory glucose, glucometer glucose, and percent reagent strip positive were lower in mucinous vs. non-mucinous cysts (P<0.0001 for all comparisons). Laboratory glucose<50 mg/dl had a sensitivity of 95% and a specificity of 57% (positive likelihood ratio (LR+) 2.19, negative likelihood ratio (LR-) 0.08). Glucometer glucose<50 mg/dl had a sensitivity of 88% and a specificity of 78% (LR+ 4.05, LR- 0.15). Reagent strip glucose had a sensitivity of 81% and a specificity of 74% (LR+ 3.10, LR- 0.26). Carcinoembryonic antigen had a sensitivity of 77% and a specificity of 83% (LR+ 4.67, LR- 0.27). The combination of having either a glucometer glucose<50 mg/dl or a carcinoembryonic antigen level>192 had a sensitivity of 100% but a low specificity of 33% (LR+ 1.50, LR- 0.00). CONCLUSIONS: Glucose, whether measured by a laboratory assay, a glucometer, or a reagent strip, is significantly lower in mucinous cysts compared with non-mucinous pancreatic cysts.