Pleural fluid glucose: A predictor of unsuccessful pleurodesis in a preselected cohort of patients with malignant pleural effusion.

PURPOSE To assess whether exclusion of patients with conditions that could lead to large fluctuations of serum glucose, would increase the accuracy of pleural fluid glucose in predicting pleurodesis outcome in patients with malignant pleural effusion subjected to bleomycin pleurodesis. METHODS A retrospective analysis of 162 patients with recurrent, symptomatic malignant pleural disease was performed. Patients with diabetes mellitus or other causes of hyperglycemia were excluded, as pleural fluid glucose has been reported to be sensitive to serum glucose fluctuations. Assessment of pleurodesis outcome was based on radiologic appearance 30 days post-bleomycin pleurodesis. RESULTS Successful pleurodesis was achieved in 64.8% of patients. Univariate analysis showed that pleural fluid glucose (p<0.001), pH (p<0.001), total proteins (p<0.001), albumin (p<0.001) and cholesterol (p<0.05) were significantly lower in patients with pleurodesis failure, while LDH was significantly higher (p<0.05). Pleural fluid glucose was the only independent predictor of pleurodesis outcome and with a cut-off point of 65 mg/dl had a high sensitivity (90.7%) with an acceptable specificity (76.8%) (p<0.001). The regression model exhibiting the highest predictive accuracy included pleural fluid glucose and albumin (sensitivity 89.3%, specificity 84.5%, p<0.001). Furthermore, a product of glucose and albumin less than 152 could predict pleurodesis failure with 88.9% sensitivity and 82.8% specificity (p<0.001). CONCLUSIONS Pleural glucose levels may reliably predict pleurodesis failure in patients without conditions that could lead to hyperglycemia, and its accuracy can increase if combined with pleural fluid albumin in an-easy-to calculate formula.